Antonellietal2021 (1)

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/353212280
Settling a family feud: a high-level phylogenomic framework for the

Gentianales based on 353 nuclear genes and partial plastomes
Article in American Journal of Botany · July 2021

DOI: 10.1002/ajb2.1697
CITATIONS READS
43 554
16 authors, including:
Alexandre Antonelli James Clarkson

Royal Botanic Gardens, Kew Royal Botanic Gardens, Kew
528 PUBLICATIONS 21,816 CITATIONS 56 PUBLICATIONS 6,988 CITATIONS
SEE PROFILE SEE PROFILE
Kent Kainulainen Olivier Maurin

Gothenburg Botanical Garden Royal Botanic Gardens, Kew
41 PUBLICATIONS 752 CITATIONS 126 PUBLICATIONS 8,133 CITATIONS
SEE PROFILE SEE PROFILE
All content following this page was uploaded by Lisa Pokorny on 13 July 2021.
The user has requested enhancement of the downloaded file.

RESEARCH ARTICLE
INVITED SPECIAL ARTICLE

For the Special Issue: Exploring Angiosperms353: a Universal Toolkit for Flowering Plant Phylogenomics
Settling a family feud: a high-level phylogenomic framework

for the Gentianales based on 353 nuclear genes and partial
plastomes
Alexandre Antonelli1,2,3,9 , James J. Clarkson1,9 , Kent Kainulainen4 , Olivier Maurin1 , Grace E. Brewer1, Aaron P. Davis1 ,
Niroshini Epitawalage , David J. Goyder
1 1
, Tatyana Livshultz , Claes Persson
5 2
, Lisa Pokorny1,6 , Shannon C. K. Straub7 , Lena Struwe8 ,
Alexandre R. Zuntini1 , Félix Forest1,* , and William J. Baker1,*
Manuscript received 6 October 2020; revision accepted 20 May 2021. PREMISE: Comprising five families that vastly differ in species richness—ranging from
1
Royal Botanic Gardens, Kew TW9 3AE, UK Gelsemiaceae with 13 species to the Rubiaceae with 13,775 species—members of the
2
Gothenburg Global Biodiversity Centre, Department of Biological Gentianales are often among the most species-rich and abundant plants in tropical
and Environmental Sciences, University of Gothenburg, Gothenburg forests. Despite considerable phylogenetic work within particular families and genera,
405 30, Sweden
several alternative topologies for family-level relationships within Gentianales have been
3
Department of Plant Sciences, University of Oxford, South Parks
presented in previous studies.
Road, Oxford OX1 3RB, UK
4
Gothenburg Botanical Garden, Carl Skottsbergs gata 22 A, METHODS: Here we present a phylogenomic analysis based on nuclear genes targeted by
Gothenburg 413 19, Sweden the Angiosperms353 probe set for approximately 150 species, representing all families
5
Department of Biodiversity Earth and Environmental Sciences and and approximately 85% of the formally recognized tribes. We were able to retrieve partial
Academy of Natural Sciences, Drexel University, 1900 Benjamin
plastomes from off-target reads for most taxa and infer phylogenetic trees for comparison
Franklin Parkway, Philadelphia, PA 19103, USA
with the nuclear-derived trees.
6
Present address: Centre for Plant Biotechnology and Genomics
(CBGP UPM –INIA), Autopista M-40, Km 38, Pozuelo de Alarcón RESULTS: We recovered high support for over 80% of all nodes. The plastid and nuclear data
(Madrid) 28223, Spain are largely in agreement, except for some weakly to moderately supported relationships.
7
Department of Biology, Hobart and William Smith Colleges, 300 We discuss the implications of our results for the order’s classification, highlighting points
Pulteney Street, Geneva, NY 14456, USA
of increased support for previously uncertain relationships. Rubiaceae is sister to a clade
8
Department of Ecology, Evolution, and Natural Resources &
comprising (Gentianaceae + Gelsemiaceae) + (Apocynaceae + Loganiaceae).
Department of Plant Biology, Rutgers University, 59 Dudley Road,
New Brunswick, NJ 08901, USA CONCLUSIONS: The higher-level phylogenetic relationships within Gentianales are
9
Authors for correspondence (e-mail: a.antonelli@kew.org, confidently resolved. In contrast to recent studies, our results support the division of
j.clarkson@kew.org)
Rubiaceae into two subfamilies: Cinchonoideae and Rubioideae. We do not formally
*Joint senior authors recognize Coptosapelteae and Luculieae within any particular subfamily but treat them
Citation: Antonelli, A., J. J. Clarkson, K. Kainulainen, O. Maurin, as incertae sedis. Our framework paves the way for further work on the phylogenetics,
G. E. Brewer, A. P. Davis, N. Epitawalage, T. Livshultz, C. Persson,
biogeography, morphological evolution, and macroecology of this important group of
L. Pokorny, S. C. K. Straub, L. Struwe, A. R. Zuntini, F. Forest, &
W. J. Baker. 2021. Settling a family feud: a high-level phylogenomic flowering plants.
framework for the Gentianales based on 353 nuclear genes and partial
plastomes. American Journal of Botany 108(7): 1–23. KEY WORDS Angiosperms353; asterids; Hyb-Seq; molecular phylogenetics; museomics;
doi:10.1002/ajb2.1697 target sequence capture; tree of life.
American Journal of Botany 108(7): 1–23, 2021; http://www.wileyonlinelibrary.com/journal/AJB © 2021 The Authors. American Journal of Botany published by
Wiley Periodicals LLC on behalf of Botanical Society of America. This is an open access article under the terms of the Creative Commons Attribution License,
which permits use, distribution and reproduction in any medium, provided the original work is properly cited. • 1
2 • American Journal of Botany
The plant order Gentianales comprises 22,510 accepted species in lobes, teeth, or crenulations) and flowers that are radially sym-
more than 1100 genera. Unrecognized species lacking a scientific metrical (actinomorphic), with an equal number of (usually few)
name are described every year, with 2019 alone seeing 157 new sepals, petals, and stamens, and a fused (tubular) corolla in which
species of Rubiaceae, second only to Orchidaceae (Cheek et al., the staminal filaments are inserted. The presence of colleters is also
2020). Some of the genera—e.g., Psychotria (Rubiaceae), with 1532 common to Gentianales (Davis and Bridson, 2007), although these
species—are larger than the great majority of plant families (Frodin, can be difficult to observe.
2004; POWO, 2020). In this study, we follow the taxonomic names Previous attempts to resolve interfamilial phylogenetic relation-
and species counts summarized in POWO (2020), where authorities ships, based primarily on plastid sequences, yielded strong support
are also provided. Ever since the first installment of the Angiosperm for the monophyly of each family, but many different topologies
Phylogeny Group classification (The Angiosperm Phylogeny Group, among the non-rubiaceous families of the order (e.g., Backlund et al.,
1998; APG IV, 2016), the order has comprised five families (by in- 2000; Frasier, 2008; Refulio-Rodriguez and Olmstead, 2014; Yang
creasing order of species richness): Gelsemiaceae, Loganiaceae, et al., 2016). Rubiaceae itself has been consistently inferred as sister to
Gentianaceae, Apocynaceae, and Rubiaceae. the remainder of the order in all but a few studies (Struwe et al., 1994;
The Gentianales have colonized all nonglaciated continental Oxelman and Bremer, 2000; Qiu et al., 2010). In contrast, studies have
land masses and are found thriving in a wide range of environments, shown contradictory and mostly weakly supported relationships
from rainforest to tundra. Fossil-calibrated molecular dating analyses among Apocynaceae, Gelsemiaceae, Gentianaceae, and Loganiaceae
suggest that the order shared a most recent common ancestor with (Fig. 2). Apocynaceae has been inferred as sister to Gelsemiaceae
the Solanales some 97–101 million years ago (Ma), whereas its ex- (Backlund, et al., 2000) (Fig. 2A), sister to Gentianaceae (Frasier,
tant lineages began to diversify 82–87 Ma (Ramírez-Barahona et al., 2008; Soltis et al., 2011; Struwe et al., 2014; Refulio-Rodriguez and
2020). The oldest reliable fossil record is Apocynospermum colora- Olmstead, 2014), and even as sister to Loganiaceae (Yang et al., 2016)
densis (Apocynaceae) from Green River, United States, the minimum (Fig. 2D). Loganiaceae and Gelsemiaceae were strongly supported as
age of which is estimated to be 48.5 Myr (Early Eocene; Brown 1929, sister families by Refulio-Rodriguez and Olmstead (2014) (Fig. 2C),
1934; Smith et al., 2003). The recently described Asclepiadospermum but formed a grade with Loganiaceae, sister to the other three fam-
spp. (Apocynaceae) of the central Qinghai-Tibetan Plateau, China are ilies elsewhere (Frasier, 2008; Soltis et al., 2011; Struwe et al., 2014)
dated to approximately the same time period (Del Rio et al., 2020). (Fig. 2B). Yang et al., (2016) produced the poorly supported topology
The order exhibits a wide morphological and ecological varia- (Gentianaceae (Gelsemiaceae (Apocynaceae + Loganiaceae))), based
tion (Fig. 1), from some minute species of high-elevation Gentiana on a taxonomically densely sampled but marker-wise sparse super-
(Gentianaceae) found in alpine habitats throughout temperate regions matrix (Fig. 2D).
to the giant Aspidosperma (Apocynaceae) trees of the Amazon, as well Phylogenomic analyses based on taxonomically sparse sam-
as epiphytic ant plants (e.g., Hydnophytum in Rubiaceae, Dischidia ples (exception: Li et al., 2019) have produced a different topol-
in Apocynaceae), and even aquatics (Limnosipanea in Rubiaceae). ogy for each genome compartment: (1) plastid (Apocynaceae
The group is also economically important: many tropical genera + Gentianaceae) + (Loganiaceae + Gelsemiaceae) (Stull et al.,
are sources of timber (e.g., Nauclea and Neonauclea in Rubiaceae, 2015, 2020; Li et al., 2019) (Fig. 2C); (2) nucleus (Apocynaceae +
Fagraea in Gentianaceae; Mark et al., 2014), and several are cultivated Loganiaceae) + (Gelsemiaceae + Gentianaceae) (Leebens- Mack
for their fruits, such as coffee (Coffea, Rubiaceae), whose annual global et al., 2019; Stull et al., 2020; Zhang et al., 2020) (Fig. 2E); and
trade is the foundation for a multibillion-dollar industry (Vegro and (3) mitochondrion (Gelsemiaceae (Gentianaceae (Apocynaceae
Florêncio de Almeida, 2020). Arguably, the two largest families of + Loganiaceae))) (Stull et al., 2020) (Fig. 2F). Interfamilial rela-
Gentianales (Rubiaceae and Apocynaceae) may not have as large tionships are strongly supported only in the nuclear topologies.
global economic trade value as some other angiosperm families—e.g., Therefore, uncertainty still exists with regards to relationships
Fabaceae, Poaceae, and Asteraceae—but they include a vast number among four of the five Gentianales families that needs to be ad-
of species with local and regional uses, particularly as medicines and dressed by a taxonomically densely sampled phylogenomic analysis.
as horticultural ornamentals. Species of Gentianales are rich sources In the following sections, we briefly outline the current phylo-
of biologically active natural products (van der Heiden et al., 2004; genetic understanding within each family and pinpoint particular
Francke et al., 2019), particularly the monoterpenoid indole alkaloids areas of uncertainty or dispute.
(found in all families, except Gentianaceae; McKey, 1980), such as the
cancer drugs vincristine and vinblastine (Cragg and Newman, 2005),
Apocynaceae
and other alkaloids such as quinine, the only effective cure against ma-
laria for centuries, first isolated from the bark of the Cinchona tree This family is easily distinguished from other Gentianales families
(Canales et al., 2020). by a combination of characters including the presence of latex, con-
Morphologically, the Gentianales can be generally identified torted corollas, and basally apocarpous gynoecia that are distally
by the presence of simple, opposite leaves with entire margins (no fused into secretory styleheads (Simões et al., 2007; Endress et al.,
FIGURE 1. A snapshot of the morphological and ecological diversity of plant order Gentianales. (A) Carissa spinarum L. (Apocynaceae). (B) Gonolobus
suberosus (L.) R.Br. (Apocynaceae). (C) Marsdenia tinctoria R.Br. (Apocynaceae). (D) Raphionacme galpinii Schltr. (Apocynaceae). (E) Mostuea brunonis
Didr. (Gelsemiaceae). (F) Anthocleista amplexicaulis Baker (Gentianaceae). (G) Chelonanthus acutangulus (Ruiz & Pav.) Gilg (Gentianaceae). (H) Gentiana
sedifolia Kunth (Gentianaceae). (I) Gentianella cerastioides (Kunth) Fabris (Gentianaceae). (J) Spigelia marilandica L. (Loganiaceae). (K) Catunaregam
oocarpa (Ridl.) Tirveng. (Rubiaceae). (L) Cinchona pubescens Vahl (Rubiaceae). (M) Galium hypocarpium (L.) Griseb. (Rubiaceae). (N) Hillia parasitica Jacq.
(Rubiaceae). (O) Palicourea weberbaueri K.Krause (Rubiaceae). (P) Razafimandimbisonia minor (Baill.) Kainul. & B.Bremer (Rubiaceae). Image credits:
Tatyana Livshultz (A–D); Kent Kainulainen (E–F, K, P); Lena Struwe (G–J ); Claes Persson (L–O).
2021, Volume 108 • Antonelli et al.—Gentianales phylogenomics • 3
A B C
D E F
G H I
J K L
M N O P
FIGURE 2. Alternative topologies for four of the five Gentianales fami-

A
GEN LOG GEL APO lies from molecular phylogenetic analysis. (A) Two plastid loci (Backlund
et al., 2000). (B) Four plastid loci (Frasier, 2008); 17 plastid, nuclear, and
mitochondrial loci (Soltis et al., 2011); three plastid loci (Struwe et al.,
2014); plastome phylogenomic (this study, Appendix S3). (C) Nine plastid
+ 1 mitochondrial loci (Refulio-Rodriguez and Olmstead, 2014); plastome
phylogenomic (Stull et al., 2015, Li et al., 2019; Stull et al., 2020). (D) Eight
plastid + 1 mitochondrial loci, sparse supermatrix (Yang et al., 2016).
(E) Nuclear phylogenomic (Leebens-Mack et al., 2019; Stull et al., 2020;
B
LOG GEL GEN APO Zhang et al., 2020; present study, Figs. 5 and 6). (F) Mitochondrial phy-
logenomic (Stull et al., 2020). Abbreviations: APO = Apocynaceae; GEL =
Gelsemiaceae; GEN = Gentianaceae; LOG = Loganiaceae. Line thickness
is approximately proportional to branch support across studies: thinnest
= unsupported, thickest = high support.
fruit and seed morphology (the Rauvolfioid grade). The rapid ra-
C diation within the APSA crown clade, which has implications for
LOG GEL GEN APO inferences of the evolution of aggregated pollen transport in the
family, has resisted conclusive resolution in the face of increasing
taxonomic and genic sampling (Livshultz et al., 2007; Livshultz,
2010; Straub et al., 2014; Fishbein et al., 2018). Likewise, rela-
tionships among the Rauvolfioid tribes Alyxieae, Melodineae,
Hunterieae, and Amsonieae have been poorly supported (Simões
et al., 2007; Fishbein et al., 2018).
D
GEN GEL LOG APO
Gelsemiaceae
This family is the smallest with three genera and 13 species.
Synapomorphies for the family are sparse, but they are all woody,
and most have dichotomously branched styles and heterostylous
yellow or white flowers (Struwe, 2018). The recent inclusion of
the enigmatic Pteleocarpa, based on molecular results, expanded
E the family’s morphological diversity (Struwe et al., 2014). Despite
GEN GEL LOG APO being a small family, understanding branching patterns within
Gelsemiaceae and its placement relative to other families will be key
to understanding the evolution of the families in the order.
Gentianaceae
F With 1840 species (Struwe and Pringle, 2018; POWO, 2020), this
GEL GEN LOG APO family is the third largest in Gentianales. The phylogenetic rela-
tionships of the gentian family have been well studied over the last
20 years, mostly using plastid markers (Struwe et al., 2002; Merckx
et al., 2013; Yang et al., 2016) and resulting in reclassification of many
genera and their tribal affiliations (Struwe et al., 2002; updated by
Struwe, 2014; Struwe and Pringle, 2018). Most people think of gen-
tians as alpine or montane herbs, but those represent highly derived
2018). It comprises ca. 5350 species classified in 378 genera grouped clades evolved from tropical ancestors of either wet or dry areas
into 25 tribes and three subfamilies (Endress et al., 2018). Twenty of (see references above). Although morphologically diverse, gentians
the tribes are not placed in subfamilies, but grouped into two infor- can often be recognized by a unified bauplan consisting of opposite
mal grades (Endress et al., 2018). entire leaves, no latex, 4–5 fused petals, and a superior ovary, often
The largest phylogenetic analyses of the family are based with parietal placentation.
on plastomes of 76 species and a 21-locus plastid supermatrix The seven main clades in the Gentianaceae are well defined
of 1045 species (Fishbein et al., 2018). The resulting topolo- and are recognized as tribes, but some relationships among them
gies support an “APSA” crown clade, comprising nine tribes are poorly supported (Struwe and Pringle, 2018). In particular, the
(the Apocynoid grade) and three subfamilies (Periplocoideae, placement of three species-rich tribes Gentianeae, Helieae, and
Secamonoideae, and Asclepiadoideae) with derived floral mor- Potalieae is uncertain and often seen as a polytomy with short and
phology and follicular fruits with comose seeds, nested in a grade poorly supported branches (von Hagen and Kadereit, 2003; Molina
of 11 tribes with plesiomorphic floral morphology and diverse and Struwe, 2009; Struwe et al., 2009). Whether this polytomy
represents a rapid diversification event or a lack of phylogenetically (visualized in their summary phylogram: fig. 1 of Wikström et al.,
informative markers remains unknown. Nearly all studies have re- 2020). Specifically, they highlighted relationships among major
lied on plastid sequences so far, and nuclear phylogenomic data is clades and the position of certain remaining problematic tribes.
needed to evaluate this further. The exact position of the achloro- Their study is based on plastomes from representatives of 56 tribes.
phyllous and mycotrophic tribe Voyrieae is also not fully settled. Overall, their inferred phylogeny was well supported with few ex-
Moreover, within Gentianaceae, there are poorly known (e.g., many ceptions, the most notable being the positions of tribes Luculieae
rare Faroinae and Helieae genera) or large polymorphic (and poly- and Coptosapelteae (previously identified as sister to all other tribes
phyletic) genera in tribes Gentianeae (Gentianella and Swertia in the family), as well as the relationships between the grade com-
in particular) and Helieae (the Calolisianthus, Irlbachia, and posed of tribes Colletoecemateae, Ophiorrhizeae, and Urophylleae
Chelonanthus complex) whose phylogenetic relationships remain and the clade comprising the rest of the subfamily Rubioideae. The
to be ascertained. authors concluded that these long-standing problems “appear to be
unaffected by the massive addition of new data” (Wikström et al.,
2020, p. 11). The equivalent ribosomal DNA dataset was also an-
Loganiaceae
alyzed in the same study and resulted in several well-supported
Before molecular studies, the Loganiaceae was a catch-all family for incongruent placements (i.e., position of tribes Anthospermeae,
asterid genera of uncertain affinities, united largely by a superior Coussareeae, Danaideae, Gaertnereae, and Jackieae), which was in-
ovary and the characters they lacked (no interpetiolar stipules, no terpreted as representing an alternative evolutionary history (i.e.,
latex, no contorted corolla in bud, among others; Struwe et al., 2018). that of ribosomal DNA).
Molecular studies reduced the size of the family drastically by show- Phylogenetic reconstructions within Rubiaceae have also been
ing that many genera should instead be placed in other families and undertaken using mitochondrial DNA. Rydin et al. (2017) presented
orders of the asterids (see Backlund et al., 2000 for an overview). At a phylogenetic hypothesis from the mitochondrial genome that was
present, 16 genera are included in Loganiaceae, with four of these largely congruent with the plastid tree of Wikström et al. (2015)
comprising the majority of its nearly 500 species; Strychnos (by far with a taxon sampling similar to that of Wikström et al. (2020).
the largest, at least 200 species), Gardneria (40 species), Mitrasacme However, they found two well-supported incongruent placements,
(55 species), and Spigelia (60–85 species; Struwe et al., 2018). when compared to the plastid DNA data: tribes Airospermeae
Owing to the multiple ethnobotanical uses of Strychnos (such as and Jackieae were supported as sister and, more surprisingly, the
for the treatment of neurological disorders, arthritis, inflammation, Cinchoneae- Isertieae clade was found nested within subfamily
and microbial infections; Patel et al., 2017), and associated detailed Ixoroideae, as sister to Posoquerieae (meaning the monophyly
floristic studies in the Neotropics, Africa, and Australia, the alpha- of subfamilies Cinchonoideae and Ixoroideae is not supported
taxonomy of Loganiaceae is relatively well-studied (summarized by by the mitome). The contradictory results obtained so far within
Leeuwenberg and Leenhouts, 1980). Recent work focused on the Rubiaceae highlight the need for further analyses and data, in par-
high species diversity of Australia clarified many relationships and ticular from the nuclear genome, which has not yet been thoroughly
led to the description of new genera (Gibbons et al., 2012). Four explored.
tribes are currently recognized, but tribe Strychneae has repeatedly
been recovered as polyphyletic, and placement of monogeneric
Aims of this study
Spigelieae is uncertain, potentially affecting the monophyly of the
other tribes. Nearly all tribes show pantropical distribution patterns. Here we assessed the higher-level relationships in the Gentianales
A detailed, full-scaled study of Loganiaceae with multiple markers using, for the first time, a taxonomically representative nuclear phy-
is still missing, even if recent work in Strychnos covers nearly half logenomic approach. We carried out concatenation and coalescent
of the species diversity in the family (Adebowale et al., 2016; R. B. analyses of 353 nuclear loci targeted by the universal sequence cap-
Setubal et al., Rio de Janeiro Botanical Garden, unpublished man- ture probe set Angiosperms353 (Johnson et al., 2019), which is de-
uscript). A reclassification of the family’s tribes is needed, as is an signed for use across all flowering plant families. We also retrieved
evaluation of complex biogeographic relationships and evolution of off-target plastid data from our sequence capture experiments,
specialized habits and life cycles (e.g., vines with tendrils, ephemeral which we used to infer a plastome phylogeny for comparison. We
herbs). asked (1) to what extent relationships inferred from our newly
generated densely sampled data set corroborate results published
to date; (2) whether the currently recognized families and subfam-
Rubiaceae
ilies in the order remain monophyletic; (3) whether relationships
The largest family of the Gentianales comprises at least 13,775 remain unclear in spite of the increased maker sampling; and (4)
species (Davis et al., 2009; POWO, 2020), with substantial annual how might these uncertainties be tackled in future studies.
increases following new scientific discoveries and taxonomic revi-
sions. Many molecular phylogenetic studies have been published
focusing on the major clades of Rubiaceae and on how they are MATERIALS AND METHODS
related. Most phylogenetic reconstructions of the family center
on three subfamilies: Rubioideae, Cinchonoideae, and Ixoroideae
Sampling
(Bremer et al., 1999; Bremer and Eriksson, 2009; Davis et al., 2009;
Rydin et al., 2009; Kainulainen et al., 2013; Mouly et al., 2014). We sequenced 143 taxa representing all five families (Apocynaceae,
Wikström et al. (2020) provided a concise summary of previ- Gelsemiaceae, Gentianaceae, Loganiaceae, and Rubiaceae) and
ous research, as well as an overview of the phylogenetic relation- all six subfamilies (three from Apocynaceae, plus the Rauvolfioid
ships of the tribes of Rubiaceae as they are currently understood and Apocynoid grades, and three from Rubiaceae). For genera
and species, we follow the classification of the Plants of the World Target Sequence Capture Kit, Angiosperms 353 v1, catalog #308196;
Online (POWO) database (http://www.plantsoftheworldonline.org, Johnson et al., 2019) following the manufacturer’s protocol (v4.0;
accessed 6 June 2020). Selected specimens were chosen to represent http://www.arborbiosci.com/mybaits-manual). Hybridizations were
the breadth of diversity within the order, including all major clades performed at 65°C for 24 h in a Hybex Microsample Incubator
according to the Angiosperm Phylogeny Group (APG IV, 2016). (SciGene, Sunnyvale, CA, USA). A volume equivalent to the hy-
Material was sourced from silica-dried samples, herbarium speci- bridization reaction (typically 30 μL) of red Chill-out Liquid Wax
mens, Kew DNA bank (http://dnabank.science.kew.org/homepage. (Bio-Rad, Hercules, CA, USA) was added on top to prevent evap-
html), and living collections at Kew (Table 1). oration. The hybridized biotin-labelled baits were then isolated us-
ing streptavidin-coated magnetic beads. Enriched products were
eluted from the beads and amplified with KAPA HiFi 2X HotStart
Molecular protocols
ReadyMix PCR Kit (Roche, Basel, Switzerland) for 8 to 12 cycles. PCR
DNA extraction—DNA was extracted from 40 mg of herbarium products were then cleaned using Agencourt AMPure XP Beads.
material, 20 mg of silica gel-dried material (Chase and Hills, 1991), Products were quantified with a Quantus fluorometer and in some
or 100 mg of fresh material using a modified CTAB extraction cases re-amplified a second time for 3 to 6 cycles. Final products
method (Doyle and Doyle, 1987). Plant tissue was pulverized using were run on a 4200 TapeStation System using D1000 ScreenTapes
a Mixer Mill MM400 (Retsch GmbH, Haan, Germany). The CTAB to assess quality and average fragment size. The enriched pools were
extraction was followed by magnetic bead clean-up using AMPure then normalized to 4–6 nM and multiplexed for sequencing to in-
XP beads (Beckman Coulter, Indianapolis, IN, USA), following the clude between 24 and 384 samples, depending on the sequencing
manufacturer’s recommendations. The bead clean-up results in a platform and service provider requirements (see below).
slight loss of yield but removes inhibitory impurities. Many existing
DNA extractions were available from the DNA bank at the Royal Sequencing—Library pools were multiplexed and sequenced on
Botanic Gardens, Kew (http://dnabank.science.kew.org/homepage. an Illumina MiSeq with v2 (300 cycles of 2 × 150-bp paired-end
html). These samples were primarily extracted using a modified reads) or v3 (600 cycles of 2 × 300-bp paired-end reads) chemistry
CTAB method (Doyle and Doyle, 1987) followed by caesium chlo- (Illumina, San Diego, CA, USA) at the Royal Botanic Gardens, Kew
ride/ethidium bromide density gradients and dialysis. or on an Illumina HiSeq producing 2 × 150-bp paired-end reads at
Genewiz (Takeley, UK) or Macrogen (Seoul, South Korea).
DNA quantification—Extracted DNA was quantified using either
a Quantus (Promega, Madison, WI, USA) or a Qubit (Thermo
Data matrix preparation
Fisher Scientific, Inchinnan, UK) fluorometer and then run in a 1%
agarose gel to assess the average fragment size. Samples with very Read mapping—The reads of the sequencing output (.fastq files)
low concentration (not visible in a 1% agarose gel), were assessed were trimmed using Trimmomatic (Bolger et al., 2014) to remove
on a 4200 TapeStation System using Genomic DNA ScreenTapes reads with a quality score below 30 and reads that had any 4-bp
(Agilent Technologies, Santa Clara, CA, USA). window below 30, retaining reads with at least 36 bp (LEADING:30
TRAILING:30 SLIDING WINDOW:4:30 MINLEN:36). The
Library preparation—DNA extracts with average fragment sizes MINLENGTH setting was also used with the length set to 36 to re-
above 350 bp were sonicated using an M220 Focused-ultrasonicator move shorter reads that might not be positioned uniquely against
with microTUBES AFA Fiber Pre-Slit Snap-Cap (Covaris, Woburn, over sequences. Paired reads and combined unpaired reads were used
MA, USA) following the manufacturer’s recommendations, with to recover target sequences using HybPiper version 1.3.1 (Johnson
shearing times of 30–90 s depending on the DNA fragment size et al., 2016) using default settings (with the exception of minimum
profile, to obtain an average fragment size of 350 bp. Dual-indexed coverage, which was set to 4×). The target file is available at GitHub
libraries for Illumina sequencing were prepared using the DNA (https://github.com/mossmatters/Angiosperms353). Reads were
NEBNext Ultra II Library Prep Kit and the NEBNext Multiplex mapped to de-gapped medoid sequences using BLASTx (Camacho
Oligos for Illumina (Dual Index Primers Sets 1 and 2) from New et al., 2009), since it has been found to result in longer sequences
England BioLabs (Ipswich, MA, USA) at either the recommended (Murphy et al., 2019). Subsequently, each gene was assembled de novo
volumes or at half these volumes. The main library preparation using SPAdes version 3.13.1 (Bankevich et al., 2012), and coding se-
steps consist of DNA repair (post-shearing), adapter ligation and quences were extracted using exonerate version 2.2 (Slater and Birney,
size selection, followed by a pre-hybridization 7–12 cycle PCR. 2005). Recovery statistics were generated using two scripts from
The Illumina indexes enable multiplexing for subsequent enrich- HybPiper v1.3 get_seq_lengths.py and hybpiper_stats.py (Johnson et
ment and sequencing. Quality of libraries was evaluated on a 4200 al., 2016), and additional statistics were prepared using custom scripts
TapeStation System using D1000 ScreenTapes ,and the libraries (available at GitHub: https://github.com/sidon ieB/). The recovery
were quantified using a Quantus fluorometer. The final average li- of plastid off-target coding sequences followed the same approach
brary size, including the adapters, was ca. 500 bp or lower when in- as for nuclear recovery, except for using BWA (instead of BLASTx),
put DNA fragments were smaller than 350 bp on average. as no significant improvement on the recovery is observed (Zuntini
et al., 2021). A plastid target file for Gentianales was produced, ex-
Target enrichment—Libraries were normalized to 10 nM, using tracting coding and ribosomal sequences from complete plastomes
10 mM Tris (pH 8.0) and then combined into pools of 20 to 24 of Loganiaceae (Mitreola yangchunensis; MT471262), Gentianaceae
libraries, each containing 10 µL (0.1 pmol) of each normalized li- (Gentiana officinalis; MH261261) and Gelsemiaceae (Gelsemium
brary (i.e., a total of ca. 600–700 ng DNA in each pool, assuming sempervirens; MG963263), as well as a plastome for each of the main
an average fragment size of ca. 450 bp). The libraries were enriched clades in Apocynaceae (Asclepias syriaca; KF386166, Hemidesmus
using the Angiosperms353 probe kit (Arbor Biosciences myBaits indicus; MN736959, Plumeria rubra; MN812495, Nerium oleander;
TABLE 1. Voucher specimens associated with sampling.

Family Genus Species Collector Collector No Museum ID
Apocynaceae Adenium obesum Chase, M.W. 727 K
Apocynaceae Allamanda cathartica [1KP] MGVU
Apocynaceae Alstonia scholaris Chantaranothai, P., Middleton, D., Parnell, J.; 911 K
Simpson, D.
Apocynaceae Alyxia buxifolia Sweedman, L., Smith, R.J.; Hopper, S.D. 202 K
Apocynaceae Amsonia hubrichtii Chase, M.W. 19252 K
Apocynaceae Apocynum venetum Prikhodko. S. 77 K
Apocynaceae Asclepias barjoniifolia Wood, J.R.I. 9532 K
Apocynaceae Aspidosperma cylindrocarpon Z
Apocynaceae Baissea sp. Cheek, M. 7533 K
Apocynaceae Catharanthus roseus [1KP] UOYN
Apocynaceae Ceropegia sandersonii Chase, M.W. 17507 K
Apocynaceae Echites umbellatus Zona, S. 758 K
Apocynaceae Eustegia minuta P. Bruyns 4357 K
Apocynaceae Fockea edulis Maurin, O. 4387 K
Apocynaceae Forsteronia viridescens M. Endress K
Apocynaceae Gonioma kamassi Chase, M.W. 5806 K
Apocynaceae Landolphia incerta Momoh, J. 134 K
Apocynaceae Lepinia taitensis S. Perlman 15071 K
Apocynaceae Ruehssia carvalhoi Chase, M.W. 17115 K
Apocynaceae Mascarenhasia lisianthiflora Brewer, G.; Maurin, O.; Rees, P. 02505 K
Apocynaceae Melodinus forbesii Takeuchi, W.; Ama, D. 17119 K
Apocynaceae Plumeria obtusa Chase, M.W. 724 K
Apocynaceae Rauvolfia vomitoria K
Apocynaceae Rhabdadenia madida Mendoza, M.; Eduardo, A.; Escobar, D. 2529 K
Apocynaceae Secamone parvifolia Goyder, D. J.; Masinde 3960 K
Apocynaceae Stephanostema stenocarpum Brewer, G.; Maurin, O.; Rees, P. 02504 K
Apocynaceae Tabernaemontana divaricata Brewer, G.; Maurin, O.; Dequiret, S. 02502 K
Apocynaceae Tacazzea venosa Dowsett-Lemairez 251 K
Apocynaceae Vallaris solanacea Chase, M.W. 3862 K
Apocynaceae Vinca difformis Chase, M.W. 9242 K
Apocynaceae Voacanga thouarsii Styles, D.G.A. 2140 K
Boraginaceae Carmona heterophylla Pullen, R. 6809 K
Boraginaceae Codon schenckii Seydel, R. 2968 K
Boraginaceae Cordia myxa Leeuwenberg, A.J.M. 7592 K
Boraginaceae Wigandia caracasana s.n. Kew 2003-2460 K
Convolvulaceae Convolvulus arvensis [1KP] CPOC
Garryaceae Garrya elliptica Chase, M.W. 1098 K
Gelsemiaceae Gelsemium sempervirens [1KP] HGSM
Gelsemiaceae Pteleocarpa lamponga Ambri & Arifin 584 K
Gentianaceae Chironia baccifera Manning, J.C.; Reeves, G. 2840 NBG
Gentianaceae Exacum exiguum Andriamahay, M.: SNGF; Rakotoarisoa, S.E. 3411 K
Gentianaceae Fagraea racemosa Johns, R.J. 9785 K
Gentianaceae Gentiana cruciata K
Gentianaceae Macrocarpaea stenophylla Weigend et al. 2000901 M
Gentianaceae Voyria aurantiaca McKey, D. 20061 K
Gesneriaceae Titanotrichum oldhamii Miyoshi Furuse 4434 K
Lamiaceae Stachys sylvatica Rahman, M.A.; Alam, M.K.; Das, S.; Hossain, Z. 1569 K
Loganiaceae Antonia ovata Fonseca, M.L.; Mendon‚àö√Öa, R.C.; Oliveira, 5913 K
F.C.A.; Pequeno, S.A.; Ara√Ä√¥jo, G.P.
Loganiaceae Geniostoma borbonicum [1KP] REU 10020 REU
Loganiaceae Logania flaviflora Brummitt, George; Oliver 21298 K
Loganiaceae Neuburgia corynocarpa Takeuchi, W.; Ama, D.; Jisaka, T. 21376 K
Loganiaceae Spigelia pulchella Harley, R.M.; Souza, V.C. s.n. K
Loganiaceae Strychnos splendens Momoh, J. 136 K
Montiniaceae Grevea eggelingii Robertson 3643 K
Oleaceae Olea europaea sp. [1KP] TORX
Rubiaceae Aidia cochinchinensis Hyland, B. 4367 K
Rubiaceae Airosperma sp. Sands, M.J.S. 6819 K
Rubiaceae Razafimandimbisonia minor Maurin, O. 100 K
Rubiaceae Aleisanthia rupestris Sands, M.J.S.; Stone, B.C. 3501 K
(Continued)
TABLE 1. (Continued)
Rubiaceae Dolianthus montiswilhelmii Brass, L.J. 29838 K
Rubiaceae Anthospermum sp. Davis, A.P. 1026 K
Rubiaceae Atractocarpus stipularis [RBG Sydney Cult] SYD
Rubiaceae Bertiera bicarpellata Davis, A.P. 3051 K
Rubiaceae Bikkia tetrandra Curry 1402 K
Rubiaceae Calycosia petiolata Barrabe, L. 1133 K
Rubiaceae Chaetostachydium barbatum Dransfield, J. 7548 K
Rubiaceae Chiococca alba Greene, L. 11 K
Rubiaceae Cinchona pubescens Maurin, O. 4385 K
Rubiaceae Coelospermum paniculatum McDonald, W.J.; Fell, D.G.; Stanton, J.P. 4525 K
Rubiaceae Colletoecema dewevrei Lisowski, S. 47195 K
Rubiaceae Coprosma longifolia Cantley, J. 120 HAW
Rubiaceae Coptosapelta diffusa Feng-Jie, Z. 256 K
Rubiaceae Cordiera concolor Chaves, E. 266 K
Rubiaceae Corynanthe pachyceras Davis, A.P. 3050 K
Rubiaceae Craterispermum sp. Davis, A.P. 2198 K
Rubiaceae Cremaspora triflora Davis, A.P. 2912 K
Rubiaceae Crossopteryx febrifugum Boane, C. 32 K
Rubiaceae Globulostylis leniochlamys Davis, A.P. 3007 K
Rubiaceae Cyanoneuron pedunculatum Rantai Jawa, J. et al. 70262 K
Rubiaceae Cyclophyllum barbatum I.Gafui & Collector BSIP 15174 K
Rubiaceae Discospermum malaccense Brun 16846 K
Rubiaceae Dolianthus vaccinioides Craven, L.A. 2790 K
Rubiaceae Dolicholobium gertrudis Takeuchi, W. 12110 K
Rubiaceae Emmenopterys henryi s.n. 001069437 (K)
Rubiaceae Gaertnera rotundifolia Maurin, O. 4397 K
Rubiaceae Galium boreale [1KP] WQRD
Rubiaceae Gardenia philastrei Davis, A.P. 4055 K
Rubiaceae Greenea corymbosa Kamarudin, S. s.n. K
Rubiaceae Guettarda pohliana Sasaki, D. 1850 K
Rubiaceae Hamelia patens Maurin, O. 4386 K
Rubiaceae Heinsia crinita Davis, A.P. 3003 K
Rubiaceae Henriquezia nitida Daly, D.C. 5480 K
Rubiaceae Hillia parasitica Forzza, R.C.; Assis, L.C.S.; Jardim, J.G.; Lima, R.; 3073 K
Menini Neto, L.; Lucas, E.; Silva, S.; Edwards, S.;
Zappi, D.
Rubiaceae Hymenodictyon floribundum Gilbert et al. 7806 K
Rubiaceae Isertia hypoleuca Vasconcelos, T. 324 K
Rubiaceae Ixora sp. Johns, R.J. 8287 BO
Rubiaceae Jackiopsis ornata Ambriansyah; Jupriansyah. 2387 K
Rubiaceae Jovetia humilis Davis, A.P. 1198 K
Rubiaceae Kajewskiella trichantha s.n. IF173 K
Rubiaceae Luculia pinceana Brewer, G.; Maurin, O.; Rees, P. 02503 K
Rubiaceae Maschalodesme sp. Takeuchi, W. 11188 K
Rubiaceae Mitragyna inermis Asase 10 GC
Rubiaceae Mitriostigma barteri Maurin, O. 6 K
Rubiaceae Mycetia sp. Maurin, O. 7 K
Rubiaceae Neolamarckia cadamba Ambriansyah; Arifin, Z. 884 K
Rubiaceae Ophiorrhiza winkleri Duaneh, J. 380 K
Rubiaceae Pachystylus zippelianus Hartley 9774 K
Rubiaceae Paederia thoursiana Maurin, O. 4388 K
Rubiaceae Palicourea nitidella Zappi, D. 856 K
Rubiaceae Pavetta ternifolia Davis 3020 K
Rubiaceae Perama dichotoma Zappi, D. 991 K
Rubiaceae Plocama calabrica Chase, M.W. 14239 K
Rubiaceae Posoqueria latifolia Maurin, O. 4390 K
Rubiaceae Prismatomeris sp. Davis, A.P. 4068 K
Rubiaceae Psychotria pandurata Davis, A.P. 2915 K
Rubiaceae Psydrax odorata Cantley, J. 111 HAW
Rubiaceae Retiniphyllum kuhlmannii Sasaki, D.; Frisby, S.; Pena, E.M.; Henicka, G.S.; 2426 K
Piva, J.H.
Rubiaceae Rondeletia buxifolia Hamilton, M. 296 K
(Continued)
TABLE 1. (Continued)
Rubiaceae Rubia peregrina Chase, M.W. 10137 K
Rubiaceae Saldinia aegialodes Davis, A.P.; Rakotonasolo, F. 2713 K
Rubiaceae Schizocolea linderi Haba, P.M. 266 K
Rubiaceae Scyphiphora hydrophyllacea Utteridge, T.M.A. 49 BO
Rubiaceae Sipanea hispida Zappi, D. 990 K
Rubiaceae Spermacoce sp. Briggs, M. 299 K
Rubiaceae Steenisia pleurocarpa Bremer, B.; Bremer, K. 1754 K
Rubiaceae Strumpfia maritima Blaise, J.; Salamanca, E.; Riggs, B.; Greenwood, K.; BNM385 K
Taylor, C.
Rubiaceae Tarenna pulchra Kesonbua, W. 157 KKU
Rubiaceae Theligonum cynocrambe Hepper, F.N. 13296 K
Rubiaceae Tricalysia verdcourtiana Mvungi, E.F. 43 K
Rubiaceae Urophyllum cyphandrum Beaman, J.H. et al. 8949 K
Rubiaceae Wendlandia sp. Davis, A.P. 4016 K
Scrophulariaceae Buddleja lindleyana [1KP] XRLM
Solanaceae Nicotiana rosulata Chase, M.W. 68280 K
Solanaceae Physalis sp. s.n. SD26 K
Solanaceae Solandra sp. s.n. SD40 K
Stilbaceae Stilbe ericoides Goldblatt, P.; Manning, J.C. 10319 NBG
Tetrameristaceae Pentamerista neotropica Kubitzki & Rohwer 91-23 HBG
Vahliaceae Vahlia capensis Chase, M.W. 5885 K
KJ953907) and in Rubiaceae (Coffea arabica; KY085909, Rubia cor- alignments. Extensive branch annotations were generated using
difolia; MN736957, Uncaria rhynchophylla; MT991006). The final ASTRAL-III version 5.5.11 (Zhang et al., 2018) and alternative quar-
target file is available as Appendix S1. tet topologies (-t 2), indicating the local posterior probabilities of the
percentage of quartets in genes. The same approach was used for the
Sequence alignments—Gene matrices were aligned sepa- plastid markers, but the results were only used to evaluate gene conflict.
rately using MAFFT version 7 (mafft- 7.419-gcc_fc6.x86), with
Accuracy-oriented methods (--localpair --maxiterate 1000) and Polytomy test—To determine whether the nuclear gene trees pro-
the option to generate reverse complement sequences to align them vided adequate information to resolve relationships in the ASTRAL
together with the remaining sequences based on 6 mer counting species tree, the polytomy test (Sayyari and Mirarab, 2018) as imple-
(--adjustdirectionaccurately). Matrices were subsequently trimmed mented in ASTRAL-III was performed (-t 10).
using phyutility (https://github.com/blackrim/phyutility) to delete
sites that are missing 80% data (-clean 0.8). Due to the fragmentary Gene concatenation (ML)—We additionally generated a phylo-
recovery of off-target plastid sequences, the alignment of plastid genetic tree using the trimmed exons alignments that were con-
markers followed the same approach as detailed above, with some catenated and simultaneously analyzed in IQ-TREE 1.6.12, as a
less stringent parameters: deleting 30% of missing data (-clean 0.3 concatenated partitioned data matrix (hereafter, nc-CON tree). Due
in phyutility) and including all genes, except ycf1 and ycf2, due to to matrix size, model selection was not feasible, so we implemented
consistent misalignment errors. the model GTR+I+G for all partitions; clade support was assessed
by 1000 replicates of UFBootstrap (BS). For the plastid data set, a
concatenated analysis was also performed (hereafter, cp-CON tree)
Tree reconstruction
now with model selection followed by partition merging to a best-
Gene trees—Gene trees from trimmed matrices were generated us- fitted partition scheme with ModelFinder in IQ-TREE 1.6.12.
ing IQ-TREE v1.6.12 (Minh et al., 2020), using ultrafast bootstrapping
with partition models (Chernomor et al., 2016). In a first iteration, Tree plotting—Trees were plotted in R (Team, 2020) using packages
we generated gene trees that were subsequently evaluated using ape (Paradis and Schliep, 2019), ggimage (Yu, 2019), ggtree (Yu et al.,
TreeShrink (Mai and Mirarab, 2018) to identify and exclude branches 2017), treeio (Wang et al., 2020), and their dependencies. In this study,
that increased the diameter of each gene tree by more than 20% with our taxonomic results and discussion are based on the final nuclear
centroid re-rooting (-b 20 -c). Each locus was then realigned, trimmed, coalescent tree generated using ASTRAL-III compared to the final
and analyzed using IQ-TREE with bipartition support. One thousand nuclear and plastid concatenation trees generated using RAxML.
ultrafast bootstrap replicates (UFBoot2) were run (-bb 1000; Hoang
et al., 2018) in IQ-TREE, and branches with support values below 10%
were collapsed (Mirarab, 2019 [Preprint]) using Newick Utilities v1.6 RESULTS
(Junier and Zdobnov, 2010). Genes that contained less than 25% of
the full sampling were excluded.
Sequence recovery
Multispecies pseudo-coalescent (ASTRAL III)—A species tree was Overall, we obtained a total of 420 billion reads, 6.2% of which
generated by inputting IQ-TREE gene tree files based on the exon were on target (i.e., angiosperms 353 loci). The percentage of
reads on target ranged from 0.24 to 18%, and the average recov- For the concatenation (Appendix S3) analysis, we report bootstrap
ery per family is as follows from lowest to highest: Gelsemiaceae, support (BS) values based on 1000 iterations. The probabilities are
3.0%; Gentianaceae, 3.2%; Rubiaceae, 4.6%; Loganiaceae, 7.3%; and based on the quartet score, that a branch is true given the set of gene
Apocynaceae, 8.0%. Over 35,000 targets (corresponding to 353 trees used to compute the species tree. We use the following three
genes), with sequences that were >75% of the target length, were terms to discuss support values in the phylogenetic trees: (1) nodes
assembled. The number of genes recovered across the sampling that with BS 100 and LPP 1 are described as fully supported; (2) nodes
were >75% of the target length was 266 (average) and 96 (median). with BS 100 and LPP > 0.85 are described as highly supported; (3)
This gene recovery, broken down per family, was as follows (aver- nodes with BS 100 and LPP 0.85 to 0.6 are described as moderately
age -median): Gelsemiaceae (65 -65), Gentianaceae (225.7 -249), supported; and (4) nodes with BS <100 and LPP <0.6 are described as
Rubiaceae (256.3 -294), Loganiaceae (295.3 -288), and Apocynaceae weakly supported. We also report cases of low and high discordance
(300.6 -310). The average read depth for the data set (exons-only) based on quartet scores and significance levels in the polytomy test
was 58×. Further statistics are provided in Appendix S1 and visual- (Fig. 5). All relationships discussed are based on our nuclear dataset
ized in Figs. 3 and 4. TreeShrink excluded 424 sequences from the (i.e., the angiosperms 353 loci) unless otherwise stated.
complete data set, and the detail per gene is provided in Appendix Both our coalescent and concatenation species trees reconstruct
S2 (see Number of Sequence pre and postTreeshrink.xlsx). The ad- Boraginales as sister to Gentianales, as inferred by other high-level
ditional filtering of genes containing less than 25% of the full sam- phylogenomic analyses (e.g., Stull et al., 2020) (Fig. 5; Appendix
pling led to four genes being excluded. S3). However, support for this relationship is moderate in our con-
catenation tree (BS 87) and rather low in our coalescent topology
(LPP 0.5; with quartet scores pointing to a high level of under-
Phylogenetic relationships
lying incongruence at the gene tree level). Gentianales are fully
We present results in relation to two metrics for statistical support of supported as monophyletic. Families are arranged into two fully
relationships, depending on the analysis. For the pseudo-coalescent supported main clades: Rubiaceae, and its sister group comprising
analysis run using ASTRAL (hereafter referred to as the coalescent the remaining families in the order. The latter four families form
analysis, see Fig. 5), we report local posterior probability (LPP) values. two fully supported clades: Gelsemiaceae + Gentianaceae, on one
A B
353
300
Number of genes
●
●
200
●
●
●
●
●
● ●
●
100 ● ●
●
Percentage of target length

●
Below 25% ●
●
●
●
Between 25% and 50%
●
●
●
Between 50% and 75% ● ●
Above 75% ●
●
● ●
● ●
●
0 ●
●
●
Samples Any length 25% or longer 50% or longer 75% or longer

Percentage of target length (total)
FIGURE 3. Number of genes recovered per sample. (A) Number of genes recovered per sample, categorised by percentage of target length. (B)
Overall number of genes per category. The percentage of the target length is colour coded (green = >75%, yellow = 50–75%, red= 25–50% and blue
= <25%).
● ●
Percentage of reads
20
●
15
on target
●
● ●
●
●
●
●
● ● ● ●
●
● ●
●
●
10 ● ●
●
●
●
●
● ● ●
● ●
●
● ● ● ● ●● ●
● ● ● ●
● ● ●
●
● ● ●
● ● ●
● ● ●
●
● ● ● ●
●
5
● ● ● ●
● ● ●● ●●
● ● ● ● ● ● ●
● ● ● ●
● ●
● ●●
● ●
● ●
● ● ●
● ●
● ● ●
● ● ● ●● ● ● ●
● ● ● ●
● ● ●
● ● ● ● ●● ● ●
● ● ●
● ● ●
● ●
0
● ●
● ●
300 ● ●
●
● ●
● ●
●
●
●
●
● ●
●
●
●
● ● ● ● ●
● ● ●
● ● ● ●
● ● ● ●● ● ● ●
● ● ●
Genes above 50%
● ● ● ● ●
●
● ● ● ● ● ● ●●●● ● ● ●
● ●
● ●
of target length
● ● ● ●
● ● ●
● ●
●
● ● ●
●
●
200
● ● ●
●● ●
●
● ● ●
●
● ●
● ● ●
● ● ● ● ● ● ●
●
● ●
●
●
●
● ● ● ●
●
●
● ●
●
●
100 ●
●
● ●
●
●
●
●
●
● ●
0
● ●
● ●
●
● ● ● ●●
● ●
● ● ●
● ● ● ● ● ●
● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ●
● ● ● ● ●● ●
● ● ● ●● ● ● ●
● ● ● ● ● ●● ● ●
● ● ● ● ● ● ● ● ● ●
● ● ●●
● ● ●● ● ●
300
● ● ● ● ●
●●
● ● ● ● ●
● ●
● ● ●
● ●
●
Number of genes
●
●
● ● ● ●
●
● ● ● ● ●
● ●
● ●
● ●
● ●
●
● ●
● ●
with exons
● ●
200 ● ●
●
●
● ●
● ●
● ●
100 ● ●
● ●
● ●
● ●
● ●
0
●
250 ● ●
● ●
● ●
200
Average exon
read depth
150
●
●
● ●
●
●
●
● ●
● ●
● ●● ●
100
●
● ●
● ●
●
● ●
●
● ● ● ●
●
● ●
● ●
● ● ● ● ●
● ●
● ● ● ●
50
● ● ● ● ●
● ● ●
● ●
● ● ● ● ●
● ●
● ● ●
● ● ●
● ● ● ● ● ● ● ●● ●● ●● ●
● ● ● ●
● ● ● ● ● ●
● ● ● ●● ●●
● ● ●
● ●
● ● ● ●
● ●
● ●● ● ● ●
● ● ● ●
● ●● ●
0
● ●
● ● ● ● ●
●
●●
Apocynaceae Gelsemiaceae Gentianaceae Loganiaceae Rubiaceae Outgroup

(n=31) (n=2) (n=6) (n=6) (n=76) (n=17)
Families
FIGURE 4. Recovery statistics for the Gentianales data set. Percentage of reads on target, genes recovered above 50% of the target length, and aver-
age read depth are assessed for exons recovered from each of the six Gentianales families.
hand, and Loganiaceae + Apocynaceae, on the other hand (Fig. 2E). Apocynaceae—All five subfamilies and 23 of 25 tribes are repre-
Monophyly of each of the five families is fully supported. sented, with full support for most relationships (Fig. 5). The co-
The plastome tree reconstructs Lamiales, not Boraginales, sis- alescent (Fig. 5) and concatenation (Appendix S3) analyses yielded
ter to Gentianales (Fig. 6). Rubiaceae is resolved as sister to the largely similar topologies, differing only in the placement of
other four families, but the topology among them shows instead Rhabdadenia madida (Rhabdadenieae); relationships among tribes
Loganiaceae, Gelsemiaceae, and Gentianaceae as successive sister Amsonieae, Hunterieae, and Melodineae; and relationships among
groups to Apocynaceae (Fig. 6; Appendices S4, S5). subtribes within Vinceae.
Maschalodesme sp Octotropideae
0.72 Jovetia humilis Octotropideae
Cremaspora triflora Octotropideae
*0.68 1.00
*0.94 Atractocarpus stipularis Gardenieae
*1.00 *0.54 Gardenia philastrei Gardenieae
Mitriostigma barteri Sherbournieae
Pachystylus zippelianus Pavetteae
*0.46 1.00 Tarenna pulchra Pavetteae
a Coffeeae
Pavetta ternifolia Pavetteae
a
1.00
Cordiera concolor Cordiereae alliance
0.56 Diplospora malaccensis Coffeeae
1.00 Tricalysia verdcourtiana Coffeeae
1.00 1.00 Bertiera bicarpellata Bertiereae
Aidia cochinchinensis Gardenieae
*0.44 1.00
Wendlandia sp Augusteae
*0.88 Razafimandimbisonia minor Alberteae
Airosperma sp Airospermeae
*0.68 *0.48 Greenea corymbosa Greeneae
Ixora sp Ixoreae
0.94 1.00 Aleisanthia rupestris Aleisanthieae
Scyphiphora hydrophyllacea Scyphiphoreae Vanguerieae
0.99
Crossopteryx febrifugum Crossopterygeae
subfam. 1.00
Globulostylis leniochlamys Vanguerieae alliance
Psydrax odorata Vanguerieae
Ixoroideae 1.00
1.00
Cyclophyllum barbatum Vanguerieae
*0.97 1.00 1.00 Jackiopsis ornata Jackieae
(previously) Retiniphyllum kuhlmannii Retiniphylleae
Heinsia crinita Mussaendeae
*0.92 Dolicholobium gertrudis Condamineeae
1.00
*0.86 Emmenopterys henryi Condamineeae
Rubiaceae
Posoqueria latifolia Posoquerieae
0.89 1.00 1.00 Henriquezia nitida Henriquezieae
Sipanea hispida Sipaneeae
Steenisia pleurocarpa Steenisieae
Bikkia tetrandra Chiococceae
1.00 Chiococca alba Chiococceae
subfam. 1.00
*0.52 0.96 Strumpfia maritima Strumpfieae
Hillia parasitica Hillieae
1.00 Hamelia patens Hamelieae
Cinchonoideae subfam. 0.98 Guettarda pohliana Guettardeae
Rondeletia buxifolia Rondeletieae
0.99 1.00
(this study) *0.65 Cinchonoideae Cinchona pubescens Cinchoneae
* 0.80 1.00 Isertia hypoleuca Isertieae
(previously) *0.43 Mitragyna inermis Naucleeae
Corynanthe pachyceras Naucleeae
1.00 Neolamarckia cadamba Naucleeae
1.00 1.00 Hymenodictyon floribundum Hymenodictyeae
Coptosapelta diffusa Coptosapelteae
Luculia pinceana Luculieae
Not classified at *0.81 Calycosia petiolata Psychotrieae
Chaetostachydium barbatum Psychotrieae
Dolianthus vaccinioides Psychotrieae
subfamily level 1.00
1.00 *0.79 Dolianthus montiswilhelmii Psychotrieae
1.00
Psychotria pandurata Psychotrieae
Palicourea nitidella Palicoureeae Psychotrieae
*0.74 1.00 Gaertnera rotundifolia Gaertnereae
Coelospermum paniculatum Morindeae
alliance
0.73 1.00 Prismatomeris sp Prismatomerideae
1.00 Craterispermum sp Craterispermeae
1.00 Schizocolea linderi Schizocoleeae
Plocama calabrica Putorieae
1.00 Theligonum cymocrambe Theligoneae
1.00 Galium boreale Rubieae
1.00 Rubia peregrina Rubieae
1.00
*0.87 1.00 Paederia thoursiana Paederieae
Anthospermum sp Anthospermeae
Coprosma longifolia Anthospermeae Spermacoceae
Cyanoneuron pedunculatum Cyanoneuroneae alliance
1.00 1.00
subfam. 1.00 1.00 Spermacoce sp Spermacoceae
Rubioideae 1.00 Mycetia sp Argostemmateae
Perama dichotoma Perameae
1.00 1.00 Saldinia aegialodes Lasiantheae
Kajewskiella trichantha Ophiorrhizeeae
1.00 Ophiorrhiza winkleri Ophiorrhizeeae
Urophyllum cyphandrum Urophylleae
0.99 *0.62 Colletoecema dewevrei Colletoecemeae
Ruehssia carvalhoi Marsdenieae
1.00 Ceropegia sandersonii Ceropegieae
1.00 Eustegia minuta Eustegieae
1.00 1.00 Asclepias barjoniifolia Asclepiadeae
Fockea edulis Fockeeae
1.00 Secamone parvifolia Secamonoideae
Apocynaceae
*0.47 1.00 Baissea sp Baisseeae
Tacazzea venosa Periplocoideae
*0.47 Vallaris solanacea Apocyneae
1.00 Apocynum venetum Apocyneae
1.00 1.00 Echites umbellatus Echiteae
1.00 Forsteronia viridescens Mesechiteae
1.00
1.00 Rhabdadenia madida Rhabdadenieae
1.00
Mascarenhasia lisianthiflora Malouetieae
Adenium obesum Nerieae
*0.89 1.00 Stephanostema stenocarpum Wrightieae
Plumeria obtusa Plumerieae
*0.52 1.00 Allamanda cathartica Plumerieae
Melodinus forbesii Melodineae
*0.86 *0.47 Gonioma kamassi Hunterieae
1.00 Amsonia hubrichtii Amsonieae
Alyxia buxifolia Alyxieae
1.00 Lepinia taitensis Alyxieae
1.00 Rauvolfia vomitoria Vinceae
0.99 Catharanthus roseus Vinceae
1.00 Vinca difformis Vinceae
0.98
Tabernaemontana divaricata Tabernaemontaneae
* 0.50
1.00
*0.97 1.00 Voacanga thouarsii Tabernaemontaneae
Landolphia incerta Willughbeieae
1.00 Alstonia scholaris Alstonieae
Aspidosperma cylindrocarpon Aspidospermateae
Geniostoma borbonicum Loganieae
0.99 1.00 Logania flaviflora Loganieae
Spigelia pulchella Spigelieae
0.99
1.00
Neuburgia corynocarpa Strychneae Loganiaceae
1.00 1.00 Strychnos splendens Strychneae
Antonia ovata Antonieae
1.00
Gentiana cruciata Gentianeae
1.00 Macrocarpaea stenophylla Helieae
1.00 Fagraea racemosa Potalieae
*0.39 *0.65 1.00 Chironia baccifera Chironieae Gentianaceae
Exacum exiguum Exaceae
1.00 Voyria aurantiaca Voyrieae
Pteleocarpa lamponga Gelsemiaceae Gelsemiaceae
1.00
1.00 Gelsemium sempervirens Gelsemiaceae Gelsemiaceae
Boraginaceae Cordia myxa
0.98 Boraginaceae Carmona heterophylla
1.00 Boraginaceae Wigandia caracasana
1.00 Boraginaceae Codon schenckii
Scrophulariaceae Buddleja lindleyana
1.00 1.00 Stilbaceae Stilbe ericoides
1.00 Lamiaceae Stachys sylvatica
1.00 Gesneriaceae Titanotrichum oldhamii
*0.59 1.00 Oleaceae Olea europaea
Vahliaceae Vahlia capensis Outgroups
NA
Solanaceae Nicotiana rosulata
1.00 Solanaceae Solandra sp
1.00 Solanaceae Physalis sp
1.00 Convolvulaceae Convolvulus arvensis
NA 1.00 Montiniaceae Grevea eggelingii
Garryaceae Garrya elliptica
Tetrameristaceae Pentamerista neotropica
FIGURE 5. Coalescent tree showing relationships in Gentianales, using nuclear exons, inferred in ASTRAL III. Local posterior probability values are
presented below branches, the families are shown in large font, and the tree tips display the genus, species, and tribe for each taxon. The pie charts
show quartet support values for each node (blue = species tree topology quartet support; red = first alternative topology quartet support; gray =
second alternative topology quartet support). Red stars next to pie charts indicate that the null hypothesis of a polytomy (equal support for species
tree, first alternative, and second alternative topology) could not be rejected with the polytomy test (p > 0.05).
Asclepiadoideae is fully supported as monophyletic and sister Amsonieae- Hunterieae form a paraphyletic grade (BS 99).
to Secamone parvifolia (Secamonoideae). Within Asclepiadoideae, Landolphia (Willughbeieae) diverges from the rest of the family af-
the tribal topology is resolved with full support, although sam- ter Aspidospermateae and Alstonieae (root to tip).
pling does not allow tests of tribal monophyly. Marsdenieae
and Ceropegieae are reconstructed as sister to each other, as are Gelsemiaceae—Since only two of the three recognized genera in
Eustegieae and Asclepiadeae, with Fockeeae as sister to the rest. the family were included in the analysis, nothing can be said about
Quartet score values indicate generally high concordance for all of intrafamilial relationships.
these relationships, with only the sister relationship of Asclepiadeae
and Eustegieae showing slightly higher discordance (Fig. 5). Gentianaceae—Six accessions, from six different tribes (tribe
Baisseeae is fully supported as sister to Asclepiadoideae + Saccifolieae is not sampled), represent the family Gentianaceae.
Secamonoideae. This clade and Tacazzea venosa (Periplocoideae) Voyrieae (mycotrophic and achlorophyllous; long branch in con-
are both nested in the Apocynoid grade, altogether forming a mono- catenation topology) is sister to a clade composed of the remaining
phyletic and fully supported APSA clade with little discordance sampled tribes. This clade, sister to Voyrieae, has moderate to low
among nuclear genes (Fig. 5). The placement of Periplocoideae support, and all other branches are fully supported. Voyria (tribe
remains uncertain and is reconstructed as sister to the aforemen- Voyrieae), Exacum (tribe Exaceae), Chironia (tribe Chironieae),
tioned Asclepiadoideae- Secamonoideae- Baisseeae clade in the and Fagraea (tribe Potalieae) are successive sister groups to
ASTRAL species tree, where high underlying discordance at the Macrocarpaea (tribe Helieae) + Gentiana (tribe Gentianeae).
gene tree level is evident from the quartet scores and a polytomy
cannot be rejected (Fig. 5, LPP 0.47), or alternatively it is placed as Loganiaceae—The relationships within this relatively small fam-
sister to Rhabdadenia madida (Rhabdadenieae) in the concatena- ily, represented by six of the 13 recognized genera, are mostly fully
tion analysis (Appendix S3, BS 100). Within the Apocynoid grade, supported and congruent between coalescent and concatenation
Apocyneae is monophyletic and fully supported as sister to a clade analyses. Antonia (tribe Antonieae) is placed as sister to all other
of Echiteae + Mesechiteae, but the placement of this clade of three tribes (LPP 1, BS 85). A monophyletic clade formed by Strychnos
tribes is uncertain in the coalescent analysis (Fig. 5, LPP 0.47, failure and Neuburgia (Strychneae) is recovered in both analyses, fully sup-
to reject null hypothesis in polytomy test). Wrightieae, Nerieae, and ported and sister to the fully supported clade of Loganieae (Logania
Malouetieae are fully supported as successive sister groups (root to and Geniostoma) plus Spigelia (Spigielieae).
tip) to the remainder of the APSA clade. However, there is moderate
discordance evident from the quartet score values for the placement Rubiaceae—We succeeded in obtaining sequences for 56 of the
of tribe Malouetieae (Fig. 5). 66 currently recognized tribes (85%). The following tribes were
The APSA clade is itself nested in the Rauvolfioid grade not sampled: Coussareeae, Danaideae, Dunnieae, Foonchewieae,
that makes up the remainder of the family. Plumerieae is highly Knoxieae, Mitchelleae, Sabiceeae, Schradereae, Seychelleeae,
supported as sister to the APSA clade, but quartet scores (Fig. 5) and Trailliaedoxeae. Four genera were included that have pre-
point to high discordance at the gene tree level. Furthermore, the viously not been included in molecular phylogenetic analyses:
polytomy test does not allow rejection of the null hypothesis that Chaetostachydium, Kajewskiella, Maschalodesme, and Pachystylus.
this node and the subtending two nodes are polytomies based on The results of the coalescent and concatenation analyses
the even distribution of quartet scores. Hence, the sister group of showed several conflicting topologies within Rubiaceae. The
the APSA clade and the placements of Plumerieae, Alyxieae, and the subfamilies Cinchonoideae, Ixoroideae, and Rubioideae are all
Melodineae-Hunterieae-Amsonieae clade remain unclear. resolved as monophyletic, although support for the clades rep-
Within the Rauvolfioid grade, Plumerieae, Alyxieae, Vinceae, resenting Cinchonoideae and Ixoroideae is moderate. In the
and Tabernaemontaneae are each fully supported as monophy- concatenation analysis, subfamily Cinchonoideae is instead para-
letic with little discordance. Clade Vinceae is highly supported as phyletic with respect to Ixoroideae, and subfamily Ixoroideae is
sister to Willughbeieae + clade Tabernaemontaneae, but with high polyphyletic because Steenisieae (subtended by a long branch) did
discordance among gene trees (Fig. 5). Within that clade, the sister not group with the rest of the subfamily—instead forming a sister
relationship of Willughbeieae and Tabernaemontaneae is equivocal group to Strumpfieae (Steenisieae formed the sister group to the
because, while the LPP is high at 0.97 in the coalescent analysis, rest of Ixoroideae in the coalescent tree). Successive sister groups
the null hypothesis of a polytomy could not be rejected, and the to the Steenisieae-Strumpfieae clade in the concatenation analy-
concatenation analysis has less than 100% bootstrap support (Fig. 5; sis are Chiococceae, a Hillieae-Hamelieae clade, a Guettardeae-
Appendix S3). The relationships among the three sampled Vinceae Rondeletieae clade, and a Cinchoneae-Isertieae clade (all nodes
subtribes is also unclear because the topology differed between are highly supported in the concatenation tree). Hymenodictyeae
analyses and high discordance among gene trees is evident in the and Naucleeae formed a clade that is moderately supported as
coalescent analysis (Fig. 5). Aspidospermateae, Alstonieae, and the the sister group to Ixoroideae in the concatenation tree, but mod-
Willughbeieae- Tabernaemontaneae- Vinceae clade are placed as erately supported as sister to the rest of Cinchonoideae in the
successive sister lineages to the rest of the Apocynaceae with full coalescent analysis (quartet score values pointing to high gene
support, although there is moderate discordance evident for the lat- tree incongruence).
ter two (Fig. 5). Within subfamily Ixoroideae, the Coffeeae alliance (repre-
A number of strongly supported clades in the plastome tree sented by the tribes Airospermeae, Alberteae, Augusteae, Bertiereae,
(Fig. 6; Appendix S5) are incongruent with strongly supported Coffeeae, Cordiereae, Gardenieae, Octotropideae, Pavetteae,
clades in the nuclear trees (Fig. 5; Appendix S3): Apocyneae and Sherbournieae) is weakly supported, and quartet score val-
is sister to Rhabdadenieae (BS 100) and Periplocoideae is ues pointed to high levels of incongruence among gene trees. In
sister to Echiteae+Mesechiteae (BS 97). Melodineae and the concatenation tree, the Airospermeae and Alberteae clade is
A Coptosapelta diffusa Tarenna pulchra B

Tarenna pulchra
Nuclear Plastid
Pachystylus zippelianus
Pachystylus zippelianus Pavetta ternifolia
Pavetta ternifolia Mitriostigma barteri
●
(IQ−TREE) ● Mitriostigma barteri
Gardenia philastrei
Gardenia philastrei
Atractocarpus stipularis
●
(IQ−TREE)
Atractocarpus stipularis Aidia cochinchinensis ●
Aidia cochinchinensis Maschalodesme sp. ●
Maschalodesme sp. ●
Jovetia humilis Jovetia humilis ●
Cremaspora triflora Cremaspora triflora ●
Discospermum malaccense Cordiera concolor
Tricalysia verdcourtiana Discospermum malaccense
Bertiera bicarpellata Tricalysia verdcourtiana ●
●
Cordiera concolor Bertiera bicarpellata ●
● Wendlandia sp. Razafimandimbisonia minor
Razafimandimbisonia minor Wendlandia sp. ●
● Airosperma sp. Airosperma sp.
Scyphiphora hydrophyllacea Ixora sp.
● Ixora sp. Scyphiphora hydrophyllacea
● Aleisanthia rupestris ●
Aleisanthia rupestris
Greenea corymbosa Greenea corymbosa ● ●
Psydrax odorata Psydrax odorata ●
Globulostylis leniochlamys
Cyclophyllum barbatum Cyclophyllum barbatum
Jackiopsis ornata Globulostylis leniochlamys ●
● Jackiopsis ornata
Crossopteryx febrifugum
Retiniphyllum kuhlmannii Crossopteryx febrifugum ●
Heinsia crinita Retiniphyllum kuhlmannii
●
Emmenopterys henryi Steenisia pleurocarpa
● Dolicholobium gertrudis Heinsia crinita ●
Sipanea hispida Emmenopterys henryi
Henriquezia nitida Dolicholobium gertrudis ● ●
●
Posoqueria latifolia Sipanea hispida
● Mitragyna inermis Henriquezia nitida ●
Corynanthe pachyceras Posoqueria latifolia
Neolamarckia cadamba Mitragyna inermis
Hymenodictyon floribundum Corynanthe pachyceras ●
Steenisia pleurocarpa ●
Strumpfia maritima Hymenodictyon floribundum ● ●
Bikkia tetrandra Neolamarckia cadamba
● Chiococca alba Strumpfia maritima
Hamelia patens Bikkia tetrandra ●
●
Hillia parasitica Chiococca alba
Guettarda pohliana Hamelia patens ●
Rondeletia buxifolia Hillia parasitica ●
● ● ●
Cinchona pubescens Cinchona pubescens
Isertia hypoleuca Guettarda pohliana
Luculia pinceana Rondeletia buxifolia
Chaetostachydium barbatum Isertia hypoleuca
● Dolianthus montiswilhelmii Luculia pinceana
Dolianthus vaccinioides Chaetostachydium barbatum
Calycosia petiolata Dolianthus montiswilhelmii ●
Psychotria pandurata ●
Palicourea nitidella Dolianthus vaccinioides ●
Gaertnera rotundifolia Calycosia petiolata ●
Prismatomeris sp. Psychotria pandurata
Cyanoneuron pedunculatum ● ●
Coelospermum paniculatum
Craterispermum sp. Palicourea nitidella
Gaertnera rotundifolia ●
Schizocolea linderi
Cyanoneuron pedunculatum Prismatomeris sp. ●
Plocama calabrica Coelospermum paniculatum
● Galium boreale Craterispermum sp. ●
Rubia peregrina Schizocolea linderi ●
Theligonum cynocrambe Plocama calabrica
● Paederia thoursiana Theligonum cynocrambe
Anthospermum sp. Rubia peregrina ●
● Coprosma longifolia Paederia thoursiana
Spermacoce sp. ●
Mycetia sp. Mycetia sp. ●
Perama dichotoma Anthospermum sp. ●
Saldinia aegialodes Coprosma longifolia
Urophyllum cyphandrum Spermacoce sp.
● Colletoecema dewevrei Perama dichotoma ●
Kajewskiella trichantha Saldinia aegialodes
Ophiorrhiza winkleri Urophyllum cyphandrum ●
Eustegia minuta Colletoecema dewevrei ●
Asclepias barjoniifolia Kajewskiella trichantha
Ceropegia sandersonii Ophiorrhiza winkleri ●
Ruehssia carvalhoi Eustegia minuta
Fockea edulis Asclepias barjoniifolia
Secamone parvifolia Ceropegia sandersonii
● Baissea sp. Ruehssia carvalhoi
Rhabdadenia madida Fockea edulis
Tacazzea venosa
Vallaris solanacea Secamone parvifolia
Apocynum venetum Baissea sp. ●
Echites umbellatus Rhabdadenia madida
Forsteronia viridescens Vallaris solanacea
Mascarenhasia lisianthiflora Apocynum venetum
Adenium obesum Tacazzea venosa
Stephanostema stenocarpum Echites umbellatus ● ●
Allamanda cathartica Forsteronia viridescens
Plumeria obtusa Mascarenhasia lisianthiflora
Gonioma kamassi Adenium obesum
Amsonia hubrichtii Stephanostema stenocarpum ●
Melodinus forbesii Plumeria obtusa
Lepinia taitensis Gonioma kamassi
Alyxia buxifolia Amsonia hubrichtii ●
Catharanthus roseus
● Vinca difformis Melodinus forbesii
Rauvolfia vomitoria Lepinia taitensis
Tabernaemontana divaricata Alyxia buxifolia ●
Voacanga thouarsii Vinca difformis
● Rauvolfia vomitoria
Landolphia incerta ● ●
Alstonia scholaris Tabernaemontana divaricata
Aspidosperma cylindrocarpon Voacanga thouarsii
Logania flaviflora Landolphia incerta
Geniostoma borbonicum Alstonia scholaris
● Spigelia pulchella Aspidosperma cylindrocarpon
Neuburgia corynocarpa Macrocarpaea stenophylla
Strychnos splendens Gentiana cruciata
●
Antonia ovata Fagraea racemosa
Macrocarpaea stenophylla Chironia baccifera ●
Gentiana cruciata
Fagraea racemosa Exacum exiguum
● Chironia baccifera Pteleocarpa lamponga
Exacum exiguum Logania flaviflora ●
Voyria aurantiaca Geniostoma borbonicum ●
Pteleocarpa lamponga Spigelia pulchella
Gelsemium sempervirens Neuburgia corynocarpa
Cordia myxa Strychnos splendens ●
Carmona heterophylla Antonia ovata
Wigandia caracasana Stilbe ericoides
Codon schenckii Stachys sylvatica ●
Stilbe ericoides Titanotrichum oldhamii
Stachys sylvatica Cordia myxa
Buddleja lindleyana Wigandia caracasana
Titanotrichum oldhamii Carmona heterophylla
● Olea europaea Codon schenckii
Vahlia capensis
● Convolvulus arvensis Vahlia capensis ●
Nicotiana rosulata Nicotiana rosulata
Solandra sp. ●
Solandra sp.
● Physalis peruviana Physalis sp.
Grevea eggelingii Grevea eggelingii
Pentamerista neotropica Pentamerista neotropica
Garrya elliptica Garrya elliptica
Apocynaceae Gentianaceae Rubiaceae 100% ● >=75% <95% ● <50%

Gelsemiaceae Loganiaceae Outgroups ● >=95% <100% ● >=50% <75%
FIGURE 6. Tanglegram showing a comparison between nuclear and plastid data. (A) Nuclear phylogeny inferred using IQ-TREE. (B) Plastid phylogeny
inferred using IQ-TREE. Nodes supports below 100% are indicated by color: green (BS ≥ 95%), yellow (75% ≥ BS >95%), red (50% ≥ BS >75%), and
black (BS <50%). Families are indicated by color of links: orange (Rubiaceae), purple (Apocynaceae), green (Loganiaceae), red (Gentianaceae), blue
(Gelsemiaceae), and gray for outgroups. See Materials and Methods for the parameters used.
weakly supported as sister to the remaining Coffeeae alliance tribes. Spermacoceae alliances is in turn sister to a clade comprising
Augusteae, Cordiereae, a Bertiereae-Coffeeae clade, Octotropideae, Perameae and Lasiantheae, all with maximum support in all trees.
Pavetteae, and Sherbournieae are resolved as a grade with respect to Colletoecemateae, Ophiorrhizeae, and Urophylleae formed a highly
Gardenieae (all nodes highly supported). In the coalescent analysis supported clade (LPP 0.99, BS 99) sister to the rest of Rubioideae
the Bertiereae-Coffeeae clade is also well supported, but most other (maximum support). In the concatenation tree, Ophiorrhizeeae is
intertribal relationships are not (quartet scores indicating high lev- highly supported as sister to Colletoecemateae, whereas Urophylleae
els of gene tree incongruence), and Gardenieae is not supported is moderately supported as sister to Colletoecemateae in the coales-
as monophyletic (Aidia not being resolved together with Gardenia cent tree (although a polytomy cannot be rejected; Fig. 5).
and Atractocarpus). Within the Psychotrieae alliance, the fully supported tribe
Unlike previous molecular phylogenetic studies based on plas- Psychotrieae formed a sister group to Palicoureeae, and this
tid data, neither the coalescent nor the concatenation analyses clade is in turn sister to Gaertnereae (all of these placements
supported a monophyletic Vanguerieae alliance (represented by with maximum support in all trees). Morindeae formed a clade to-
the tribes Aleisanthieae, Crossopterygeae, Greeneeae, Ixoreae, gether with Prismatomerideae. Relationships between the Morindeae-
Jackieae, Scyphiphoreae, and Vanguerieae), which is instead a Prismatomerideae clade, the Psychotrieae-Palicoureeae-Gaertnereae
grade leading to the Coffeeae alliance clade. In the concatenation clade and the tribes Craterispermeae and Schizocoleeae are moder-
tree Aleisanthieae, Greeneae, Ixoreae, and Scyphiphoreae formed ately supported in the coalescent analysis (probably resulting from
a highly supported clade sister to the Coffeeae alliance with weak underlying gene tree incongruence, as evidenced by quartet score val-
support, whereas a Crossopterygeae-Jackieae-Vanguerieae clade is ues). Relationships within tribe Psychotrieae receive moderate sup-
inferred as sister to the clade composed of the tribes listed in the port, and a polytomy cannot be rejected (Fig. 5). In the concatenation
previous sentence plus all tribes in the Coffeeae alliance with weak tree Craterispermeae and Schizocoleeae are resolved, with high sup-
support. Backbone branch lengths in this part of the concatenation port, in a grade successively sister to the aforementioned tribes in the
tree are very short, whereas branches subtending tips are rather Psychotria alliance. Most clades within the Spermacoceae alliance are
long (which could indicate both long-branch attraction and incom- well supported in both analyses. In the coalescent tree, Putorieae and
plete lineage sorting are at work). In the coalescent analysis, most Theligoneae formed a highly supported clade that is sister to Rubieae.
of these intertribal relationships are poorly supported (plus quar- Successive sister groups to the Putorieae- Theligoneae-Rubieae
tet scores point to very high gene tree incongruence), although the clade are in turn Paederieae, Anthospermeae, Cyanoneuroneae,
Aleisanthieae-Greeneae-Ixoreae clade and Jackieae-Vanguerieae Spermacoceae, and Argostemmateae. Clade support is high ex-
clade are resolved with high support. cept for the early-diverging clades. Except for the highly supported
Both the coalescent and the concatenation analyses resolved Paederieae-Putorieae-Theligoneae-Rubieae clade, the concatenation
Retiniphylleae as sister to the clade comprising the Coffeeae alliance analysis found only weak support for intertribal relationships within
clade and the paraphyletic Vanguerieae alliance grade. This clade the Spermacoceae alliance.
is in turn sister to Mussaendeae. Condamineeae, Henriquezieae,
Posoquerieae, and Sipaneeae form a fully supported clade sister to
all other Ixoroideae, Steenisieae excepted. Within this latter clade, DISCUSSION
Condamineeae is weakly supported as sister to a Henriquezieae-
Posoquerieae clade in the coalescent analysis, and this clade is in The interfamilial relationships presented here derive from the most
turn highly supported as sister to Sipaneeae. In the concatenation densely sampled phylogenomic analysis of Gentianales to date. Our
tree Sipaneeae is instead resolved as sister to Condamineeae. target sequence capture approach yielded an extensive nuclear data-
Luculieae and Coptosapelteae are in a grade leading to a highly set, comprising over 35,000 targets with sequences that were >75%
supported Cinchonoideae-Ixoroideae clade, in both the coalescent of the target length. As predicted, the percentage of reads on target
and concatenation analyses, all of them forming a fully supported is relatively low (6.2%) using the universal Angiosperms353 probe
clade. Coptosapelteae is resolved as sister to the Cinchonoideae- set, but this shortcoming is offset by the time and money saved by
Ixoroideae clade, a relationship that is weakly supported, and a not having to develop a custom probe set. Also, these data can be
polytomy cannot be rejected (Fig. 5). used in conjunction with data from other groups to form much
Subfamily Rubioideae is highly supported as monophyletic in larger phylogenetic trees spanning the breadth of all angiosperms
both the coalescent (quartet score values showed high congruence (see Baker et al., 2021). Despite the fairly low percentage on target,
of underlying gene trees) and the concatenation analyses. Within we were able to recover an average of 266 and median of 296 genes
Rubioideae, the fully supported Psychotrieae alliance (represented by out of 353 genes. The factors affecting target sequence capture ef-
the tribes Craterispermeae, Gaertnereae, Morindeae, Palicoureeae, ficiency have previously been explored using Rubiaceae as a focus
Prismatomerideae, Psychotrieae, and Schizocoleeae) is well sup- family (Brewer et al., 2019); therefore, we refrain from a detailed
ported as the sister group to the fully supported Spermacoceae al- appraisal of methodological aspects here and focus on key phylo-
liance (represented by the tribes Anthospermeae, Argostemmateae, genetic insights.
Cyanoneuroneae, Paederieae, Putorieae, Rubieae, Spermacoceae, As in most previous analyses (Fig. 2), Rubiaceae is fully supported
and Theligoneae). The clade formed by the Psychotrieae and as the sister to the other four families in the order. Relationships
among the other four families, (Gentianaceae + Gelsemiaceae) + radiation; within that crown clade, Baisseeae (tropical Africa and
(Apocynaceae + Loganiaceae), are also fully supported, confirming Madagascar) and Secamonoideae (paleotropics) as successive sis-
the results of previous, taxon-sparse, nuclear phylogenomic anal- ter groups to Asclepiadoideae (global); and relationships among
yses (Fig. 2E) (Leebens-Mack et al., 2019; Stull et al., 2020; Zhang Asclepiadoideae tribes (see below for details). High congruence
et al., 2020). In contrast, the plastome data set (Fig. 2; Appendices among many unlinked loci indicates that these relationships likely
S4 and S5) yields one of the four topologies reported in previ- depict the evolutionary history of these lineages and that infer-
ous plastid- based studies (Fig. 2B) placing Gentianaceae and ences about morphological evolution and biogeography drawn
Apocynaceae as sister taxa with high support (BS 97/99), but pro- from previously published topologies remain valid (Fishbein et al.,
viding poor support for Loganiaceae as sister to the other three (BS 2018 and references therein). However, many outstanding ques-
0/88). Cytonuclear discordance has been identified elsewhere in the tions about tribal relationships in the family pertain to monophyly
radiation of Asteridae (Stull et al., 2020) and linked to ancient poly- of tribes that are here represented by only a single species each
ploidization events. Polyploidy has not been inferred in the early di- (e.g., Willughbeieae, Melodineae, Nerieae) or are not sampled (i.e.,
versification of Gentianales (Leebens-Mack et al., 2019; Zhang et al., Odontadenieae) (Endress et al., 2018; Fishbein et al., 2018).
2020), but homoploid hybridization and lineage sorting are possible
causes of the observed incongruence that need to be investigated.
Apocyneae, Echiteae-Mesechiteae, APSA crown radiation
The consistent support for the nuclear topology from data sets with
varying taxon and gene sampling suggests that it represents the his- The only novel relationship that received high, consistent support from
tory of most of the genome. nuclear loci is the monophyly of a clade including Apocyneae (an al-
Concatenation analyses are based on combining alignments most exclusively tropical Asian tribe) and Echiteae plus Mesechiteae
to form a “supermatrix” used to directly infer a species tree (i.e., (an almost exclusively tropical American lineage). These two lineages
all genes enforced to the same history, varying or not the substi- are part of the APSA crown radiation that had resisted resolution
tution models); organellar genomes are good targets for these with small amounts of data (Livshultz, 2010). In plastome topologies
approaches as all markers are linked. Multispecies coalescent (Straub et al., 2014, Fishbein et al., 2018; Fig. 6), Apocyneae is sister to
analyses (MSC), on the other hand, are based on individual gene Rhabdadenieae (tropical America) while Echiteae, Mesechiteae, and
trees that are summarised into a species tree (Jiang et al., 2020). the here unsampled Odontadenieae (tropical America) form a clade
This latter method is capable of tackling challenges on phyloge- sister to Periplocoideae (Paleotropics). Thus, the novel support for
netic reconstruction, such as incomplete lineage sorting and long- Apocyneae + Echiteae-Mesechiteae is due to the phylogenetic signal
branch attraction. Since the premise of MSC is unlinked markers, from the nuclear loci rather than a difference in taxon sampling com-
it is mostly used for nuclear markers, although it might be used for pared to prior plastid analyses.
organellar data, to evaluate conflict among genes (Pérez-Escobar Straub et al. (2014) showed that the plastome data set was ac-
et al., 2021; Zuntini et al, 2021). Although these two approaches are tually ambivalent among a subset of possible resolutions of the
fundamentally different, they often result in similar topologies for crown clade radiation and that discrimination increased with re-
data sets where the underlying signal is strong, as demonstrated by moval of rapidly evolving sites. Their analysis was focused on the
our results (see comparison in Appendix S6) and observed in others position of Periplocoideae, but a similar analysis can be conducted
groups such as Apiaceae (Clarkson et al., 2021), Myrtales (Maurin to evaluate whether the plastome data actually reject the recov-
et al., 2021), Ochnaceae (Shah et al., 2021), and Sapindaceae (Buerki ery of a clade comprising Apocyneae, Echiteae, Mesechiteae, and
et al., 2021). Odontadenieae. In any case, the consistent support for Apocyneae-
Echiteae-Mesechiteae by multiple unlinked loci in the nuclear anal-
ysis provides compelling evidence that this topology represents the
Apocynaceae
evolutionary history of most of the genome, while the plastid topol-
The nuclear topology presented here (Fig. 5, Appendix S3) is similar ogy resulted from other processes, such as lineage sorting or plastid
to a previously published plastome topology (Fishbein et al., 2018), capture (Wendel and Doyle, 1998).
as well as topologies from much smaller plastid data sets (Livshultz Different placements of Rhabdadenieae and Periplocoideae are
et al., 2007; Simões et al., 2007). Of the 29 clades in the Apocynaceae resolved in the concatenation (Appendix S3) and coalescent analy-
subtree (Fig. 5), 23 are also identifiable in the plastome tree of ses (Fig. 5), and none of these match positions in plastome data sets
Fishbein et al. (2018). The new plastome topology (Fig. 6) is com- (Fig. 6). Short internal branches and long terminal branches, discor-
pletely congruent with that of Fishbein et al. (2018), except for the dance among unlinked genetic markers, continued topological un-
position of Willughbeieae, which Fishbein et al. (2018) recovered in certainty and low support with increasing sequence sampling are all
a strongly supported clade with Vinceae and Tabernaemontaneae hallmarks of ancient rapid radiations (Rokas and Carroll, 2006). The
(as we recover in our nuclear analyses [Fig. 5, Appendix S3]). We polytomy tests show that there is no statistically significant support
suggest the aberrant position of Willughbeieae in our plastome for the species tree resolution (Fig. 5) over the two alternative res-
analysis is an artifact of missing data (only 14 genes present). olutions of the quartet topology in the placement of Rhabdadenia
Relationships receiving high support from the nuclear genes (Rhabdadenieae) and Tacazzea (Periplocoideae). The branches sub-
(Fig. 5; Appendix S3) and the plastome (Fishbein et al., 2018) in- tending Rhabdadenieae and Periplocoideae can be shortened by in-
clude: Aspidospermateae (tropical and temperate America), creasing taxonomic sampling, which may decrease uncertainty. But
Alstonieae (paleotropics), and a Vinceae- Tabernaemontaneae- we may also have to accept that the position of these two taxa in
Willughbeieae clade (paleo-and neotropics) as successive (root to the crown clade topology will remain ambiguous and use a sample
tips) sister groups to the rest of the family; the APSA clade with of equally likely topologies to make inferences about evolution and
Wrightieae (paleotropics), Nerieae (paleotropics), and Malouetieae homology of aggregated pollen and other floral novelties shared by
(paleo-and neotropics) as successive sister groups to a crown Periplocoideae (pollen in tetrads or pollinia) and Asclepiadoideae
plus Secamonoideae (pollen in pollinia) (Straub et al., 2014). by whole plastome data (Straub et al., 2013; Fishbein et al., 2018;
Importantly, Baisseeae (pollen in monads) is consistently supported Fig. 6), should now be considered firmly established. It is also in
as sister to Asclepiadoideae and Secamonoideae, which means that line with morphology, as pendant pollinia are observed in both
there must have been more than one evolutionary transition be- tribes (Bruyns, 1999).
tween monads and aggregated pollen.
Loganiaceae
Alyxieae, Melodineae-Hunterieae-Amsonieae, Plumerieae
The pantropical tribe Antonieae (4 genera, one to a few species in
The branches placing these three lineages as successively (root each) has been repeatedly found to be sister to the rest of the fam-
to tips) sister to the APSA clade all have nonsignificant poly- ily and is supported in our nuclear concatenation analysis. Tribe
tomy tests, indicating a lack of support for the topology. In plas- Strychneae in its current circumscription includes three genera, the
tid topologies, the position of Plumerieae (tropical America well-known and speciose Strychnos and the smaller Asian-Pacific
and Pacific) has been unambiguously sister to a clade including Neuburgia and Gardneria. The monophyly of tribe Strychneae has
Carisseae (Paleotropics, here unsampled) and the APSA clade been questioned repeatedly due to the position of Gardneria in a
(Simões et al., 2007; Fishbein et al., 2018). It remains to be de- different clade from Strychnos and Neuburgia (R. B. Setubal et al.,
termined whether the absence of Carisseae is the cause of un- Rio de Janeiro Botanical Garden, unpublished manuscript), but
certainty in the position of Plumerieae in the present analyses. this monophyly could not be tested here because Gardneria was not
Monophyly of the clade composed of Melodineae (tropical Africa sampled.
and Asia), Hunterieae (tropical Africa), and Amsonsieae (tem- A highly supported relationship between Geniostoma and
perate Eurasia and America) is highly supported by the nuclear Logania (tribe Loganieae) has been found in several analyses
data set with a significant polytomy test, but the sister relation- (e.g., Gibbons et al., 2012), while the exact relationships of tribe
ship of Melodineae and Hunterieae is not. Relationships among Spigeliaeae (only including Spigelia, mostly neotropical) are not
these tribes, as well as Alyxieae, were not well supported in pre- clear yet. An expanded sampling within Loganiaceae would ad-
vious plastid-derived topologies (Simões et al., 2007; Fishbein dress many evolutionary questions related to pantropical bio-
et al., 2018), and the sister relationship of Melodineae to the geographic patterns (three tribes have pantropical distributions),
Hunterieae-Amsonieae clade is contradicted (Fig. 6). This part evolution of fruit types, ovary fusion, and placentation (highly di-
of the tree shows the characteristic short internal branches and verse in the family), woodiness, and habit (from ephemeral herbs
long terminal branches of an ancient rapid radiation (Fishbein to large trees and lianas), as related to environments and past
et al., 2018). Furthermore, while Hunterieae (four genera) and climates.
the monogeneric Amsonieae are each highly supported as mono- It would be of great interest to compare the traits inferred
phyletic (Simões et al., 2007), Melodineae is resolved as polyphy- as ancestral in Loganiaceae to traits identified as ancestral in
letic with Diplorrhynchus sister to Alyxieae and Pychnobotrya Apocynaceae (Simões et al., 2007; Fishbein et al., 2018). These two
sister to the Vinceae-Tabernaemontaneae-Willughbeieae clade; sister families have evolved in parallel but, while Loganiaceae still
only Melodinus and Craspidospermum formed a monophyletic has the general bauplan of 4–5-merous asterid flowers, key innova-
clade (Fishbein et al., 2018). Including all genera of Melodineae tions in the floral morphology within Apocynaceae might have led
is key to resolving both the continued uncertainty regarding the to its more rapid radiation. The plants of the two families are often
circumscription of this tribe and the intertribal relationships in found in similar tropical habitats, from dry savannas to rainforests.
this group. It is notable that much of the species diversity in the Apocynaceae is
in the Asclepiadoideae where the waxy coating of the pollinia may
Asclepiadoideae tribes have enabled radiations into seasonally dry environments and
drought-stress microhabitats such as epiphytes within more humid
In all analyses, the highly supported relationships among the five vegetation (Livshultz et al., 2011).
tribes of Asclepiadoideae, with Fockeeae (Africa) sister to a clade
composed of the other four tribes, where Asclepiadeae (cosmo- Gentianaceae
politan) and Eustegieae (southern Africa) form a clade that is
sister to another clade comprising Marsdenieae (widespread) Only six species representing six of the seven tribes (with over 1700
and Ceropegieae (Old World), are identical to those recovered species total) of Gentianaceae are included in this analysis, but the
previously in phylogenetic analyses of plastid data (Fishbein inferred relationships corroborate the results from larger, recent
et al., 2018). While the topological relationships of Fockeeae, analyses (Struwe and Pringle, 2018, and references therein). There
Asclepiadeae, Marsdenieae, and Ceropegieae have been con- is now a good understanding of the phylogeny of Gentianaceae,
sistently recovered by molecular phylogenetics studies (Liede, with the exception of some rare and yet to be sequenced genera
2001; Liede and Täuber, 2002; Rapini et al., 2003, 2007; Liede- and, maybe more importantly, the relationship among the tribes
Schumann et al., 2005, 2014; Straub et al., 2013; Surveswaran Gentianeae (primarily temperate, cosmopolitan), Helieae (neo-
et al., 2014), Eustegieae, with two monotypic genera, is alterna- tropical), and Potalieae (pantropical). There have been suggestions
tively resolved as either sister to Ceropegieae plus Marsdenieae that a rapid radiation took place in the early evolutionary history of
(Rapini et al., 2003, 2007; Meve and Liede, 2004) or as sister to Helieae and Potalieae, since internal branch lengths appeared short
Asclepiadeae (Liede, 2001; Surveswaran et al., 2014). The sister and poorly supported. It is clear that Voyria, the only genus of tribe
relationship of Eustegieae and Asclepiadeae, that is highly sup- Voyrieae, is a gentian, but its relationship to tribes Saccifolieae and
ported here by the nuclear gene data and is also highly supported Exaceae at the base of the family is still uncertain.
Rubiaceae The branch lengths are short in this part of the nuclear tree, yet
Colletoecemateae and Ophiorrhizeae are highly supported as sis-
The phylogeny of Rubiaceae inferred in the coalescent analysis is
ters in the concatenation tree, which is not the case in the coalescent
largely congruent with that found in previous molecular phyloge-
analysis.
netic studies based on plastid DNA data (e.g., Wikström et al., 2015,
2020).
Cinchonoideae grade—In previous molecular studies, resolv-
ing relationships among low- diversity tribes within subfamily
Coptosapelteae and Luculieae—The phylogenetic positions of Cinchonoideae has proved difficult (e.g., Rova et al., 2002; Bremer
the SE Asian tribes Coptosapelteae and Luculieae have been a long- and Eriksson, 2009) or has resulted in alternative topologies, where
standing problem in Rubiaceae systematics, and they remain un- either the Hymenodictyeae-Naucleeae clade or the Cinchoneae-
classified at subfamily level (Wikström et al., 2020). These tribes Isertieae clade are sister to all other Cinchonideae tribes (e.g.,
were emended and described, respectively, by Rydin and Bremer Andersson and Antonelli, 2005 vs. Manns and Bremer, 2010).
(2009), who found that Coptosapelteae (comprising Acranthera Wikström et al. (2020, p. 11) found incongruent topologies in analy-
and Coptosapelta), and the monogeneric Luculieae, did not ses of the plastid genome depending on whether DNA data was an-
group together with the subfamilies Cinchonoideae, Ixoroideae, alyzed at the nucleotide (Hymenodictyeae-Naucleeae clade sister)
or Rubioideae in an analysis of combined plastid and nrITS data. or at the codon level (Cinchoneae-Isertieae clade sister), and they
Instead, a Coptosapelteae-Luculieae clade is resolved but with low concluded that this apparent incongruence was “likely resulting
support (from the nrITS data). Wikström et al. (2020) also found from erroneous assumptions in the models used to reconstruct our
support for a Coptosapelteae-Luculieae clade in their analysis of ri- phylogenies”. Another surprising result was found in the analyses
bosomal DNA, but only when the data is analyzed using a Bayesian of the mitochondrial genome, by Rydin et al. (2017), which showed
Markov chain Monte Carlo method (the clade is not supported in that the Cinchoneae-Isertieae clade did not group together with the
their maximum likelihood analysis). When analyzing combined rest of the subfamily, instead being nested inside Ixoroideae. The
coding and noncoding data from the plastid genome, the authors authors hypothesized that ancient hybridization could explain why
instead found that Luculieae formed a well-supported clade with the organellar genomes yield this conflicting phylogenetic signal. In
Rubioideae, whereas the position of Coptosapelteae (sister group to the present study, the coalescent analyses resolve Cinchonoideae as
the Luculieae-Rubioideae clade) was not highly supported. a monophyletic but moderately supported clade, whereas the results
In contrast, in the phylogeny presented by Rydin et al. (2017) of the concatenation analysis indicate that the subfamily is paraphy-
and based on mitochondrial data, Coptosapelteae was highly sup- letic with respect to subfamily Ixoroideae.
ported as sister to the Cinchonoideae-Ixoroideae clade (the position
of Luculieae was poorly supported as sister to Rubioideae). Their Tribe Steenisieae—In the coalescent and concatenation analyses of
result, in which Coptosapelteae and Luculieae formed a highly sup- this study, the monogeneric tribe Steenisieae is resolved in highly
ported clade together with the Cinchonoideae-Ixoroideae clade, is supported but conflicting positions within Rubiaceae. In the coales-
the most congruent with what we find in this study. A Coptosapelteae- cent analysis, it is sister to the rest of subfamily Ixoroideae, whereas
Luculieae clade is not supported in our analyses, instead Luculieae the concatenation analysis finds it subtended by a long branch and
and Coptosapelteae are in a (root to tip) grade leading to the nested within Cinchonoideae, as sister to Strumpfieae. Neither result
Cinchonoideae-Ixoroideae clade. This relationship is poorly sup- is congruent with previous molecular phylogenetic studies that have
ported (LPP 0.65; BS 82), and the null hypothesis of the polytomy supported an isolated position within Ixoroideae, as sister group to
test could not be rejected for the inferred topology (Fig. 5). However, a clade comprising the Coffeeae alliance, the Vanguerieae alliance,
it should be noted that the sequence recovery for the Coptosapelta and Retiniphylleae (e.g., Kainulainen et al., 2013; Rydin et al., 2017;
sample was the second least successful of all samples (Appendix S1). Wikström et al., 2020). Steenisieae and Strumpfieae have widely dis-
The evolutionary history of Coptosapelteae and Luculieae ap- junct distributions (restricted to SE Asia and the Caribbean, respec-
pears complex. In terms of their classification, there are at least three tively), so a close relationship is surprising from a biogeographic
options: (1) recognizing the lineages represented by Coptosapelteae point of view. However, it is notable that they both share a morpho-
and Luculieae as two separate subfamilies; (2) including them in logical trait that is very uncommon in the family: fused anthers that
Cinchonoideae sensu lato; or (3) not formally recognizing them un- form a cone around the style (see Puff et al., 1995).
der any particular subfamily, i.e. as incertae sedis. We discuss these
options below. Paraphyly of the Vanguerieae alliance—Our study does not support
the monophyly of the informally named Vanguerieae alliance, a clade
Rubioideae grade—Phylogenetic relationships among the species- that includes the tribes Aleisanthieae, Crossopterygeae, Greeneeae,
poor clades within subfamily Rubioideae outside the Psychotrieae Ixoreae, Jackieae, Scyphiphoreae, Trailliaeodoxeae, and Vanguerieae
and Spermacoceae alliances have also been difficult to resolve, of- and which has been, in general, highly supported in plastid DNA anal-
ten lacking satisfactory support or showing conflicting topologies. yses (Razafimandimbison et al., 2011; Wikström et al., 2015, 2020).
Depending on the data analyzed, either tribe Colletoecemateae Two clades are instead resolved in the concatenation analysis, one
(Rydin et al., 2009), a Colletoecemateae-Urophylleae clade (Rydin comprising Crossopterygeae, Jackieae, and Vanguerieae, and the other
et al., 2017), or a Colletoecemateae- Ophiorrhizeae- Urophylleae Aleisanthieae, Greeneeae, Ixoreae, and Scyphiphoreae. These clades
clade (Wikström et al., 2020) have been inferred as sister to the re- form a grade with respect to the Coffeeae alliance clade, with the
maining Rubioideae. In this study, we find high support for a clade Crossopterygeae-Jackieae-Vanguerieae clade sister to all other tribes.
comprising Colletoecemateae, Ophiorrhizeae, and Urophylleae that The estimated backbone branch lengths are very short in this part
is sister to the rest of the subfamily. This relationship is congruent of the concatenation tree, suggesting a rapid radiation. It should be
with the results of Wikström et al. (2020) based on nrDNA data. noted that phylogenetic reconstructions based on the mitochondrial
genome support yet another topology, in which Jackieae forms a clade a mono-tribal Guettardoideae (later changed to Antirheoideae in
with Airospermeae (see Rydin et al., 2017). Verdcourt [1989]). The Cinchonoideae and Rubioideae were char-
acterized by the absence vs. presence of raphides, respectively, while
Phylogenetic positions of newly sequenced genera—Kajewskiella is the Guettardoideae was characterized by a large embryo, strongly
a genus endemic to the Solomon Islands that has not been included reduced endosperm and no raphides, although there are exceptions
in previous molecular phylogenetic studies. In the original descrip- (Davis and Bridson, 2007). The placement of tribe Guettardeae
tion of Kajewskiella, Merrill and Perry (1947, p. 332) were “unable with members of Cinchonoideae s.l., and the support for other
to suggest an affinity” for their new genus, which they compared to higher-level relationships, led Robbrecht and Manen (2006) to ad-
Bikkia Reinw. ex Blume, Dolicholobium A.Gray, and Xanthophytum vocate a two-subfamily classification for the Rubiaceae.
Reinw. ex Blume. These genera are currently assigned to the tribes Taking into account our findings and the results presented
Chiococceae, Condamineeae, and Ophiorrhizeae, respectively. Jansen by Rydin et al. (2017) and Wikström et al. (2020), we pro-
(1979), tentatively included Kajewskiella in Condamineeae, whereas pose that the Rubiaceae should comprise the two subfamilies
Tange (1995) and more recently Razafimandimbison and Rydin Cinchonoideae and Rubioideae, in agreement with Robbrecht
(2019) considered Kajewskiella to be closely related to Xanthophytum and Manen (2006). Our results indicate that Coptosapelteae
and part of Ophiorrhizeae. The results of this study support the latter and Luculieae are most closely related to Cinchonoideae; their
classification because we find Kajewskiella resolved in a highly sup- inclusion in that subfamily would lead to difficulties in charac-
ported clade together with Ophiorrhiza L. in all our trees. terizing the expanded clade morphologically, particularly con-
Chaetostachydium, which was originally described under the sidering their unique morphological features (Robbrecht, 1988;
illegitimate name Chaetostachys Valeton, comprises three spe- Rydin et al., 2009). We therefore propose that Coptosapelteae and
cies of pachycaul treelets from New Guinea. A taxonomic revi- Luculieae not be formally recognized within any particular sub-
sion was done by Ridsdale (1975, p. 269), who expressed doubts family but rather treated as incertae sedis, until conclusive evi-
that Chaetostachydium is “sufficiently distinct from Psychotria L. dence is presented.
to warrant generic distinction”, based on morphological evidence.
In this present study, we find Chaetostachydium nested within the
tribe Psychotrieae in a clade together with Amaracarpus Blume, CONCLUSIONS
Calycosia A.Gray, and Dolianthus C.H.Wright. Razafimandimbison
et al. (2014) recently performed molecular phylogenetic analy- Our study constitutes the largest phylogenomic effort to date to re-
ses of Psychotrieae and concluded that Psychotria is paraphyletic solve the high-level relationships in the plant order Gentianales. By
and better circumscribed to include the entire tribe, e.g., including generating, processing, and analyzing over 420 billion DNA reads of
Amaracarpus, Calycosia, and Dolianthus. Chaetostachydium was not the nuclear genome from nearly 150 species, we were able to obtain
sampled in their study, but given its close relationship in our analysis high phylogenetic support for the vast majority of clades inferred.
to Amaracarpus, Dolianthus, and the remaining species of Calycosia, Perhaps most noteworthy, the combination of dense taxonomic sam-
it seems that this genus should also be included in Psychotria. pling and our nuclear phylogenomic analyses provide an unambig-
The two species of the small genus Maschalodesme are also uous answer to a long-standing debate on the relationships among
pachycaul treelets from New Guinea. On morphological grounds, the five families of the order. Our results confirm that the Rubiaceae
Ridsdale et al. (1972, p. 346) compared it to a few genera (“allies are sister to all other families, that Gelsemiaceae are sister to
of Diplospora”) that are currently assigned to the tribes Coffeeae, Gentianaceae, and that Apocynaceae are sister to Loganiaceae, thus
Gardenieae, or Octotropideae. In a morphological study of building the sequence: (Rubiaceae ((Gentianaceae, Gelsemiaceae)
Gardenieae and related tribes, Robbrecht and Puff (1986) placed (Apocynaceae, Loganiaceae))).
Maschalodesme in “Hypobathreae” (=Octotropideae). This clas- Ever since the seminal phylogenetic studies of angiosperms using
sification is supported by the results presented here as we find single-locus data sets (e.g., Chase et al., 1993) the plant systematic
Maschalodesme nested within Octotropideae. community has gradually moved toward harnessing increasingly
Pachystylus comprises two species from Melanesia and is con- larger genomic data. Recent advances in high-throughput DNA se-
sidered to be a member of tribe Pavetteae and very close or conge- quencing have provided a step-change in data acquisition, but they
neric to Tarenna Gaertn. (Bridson and Robbrecht, 1985; De Block have been accompanied by a lack of international coordination in
et al., 2015). Our study supports this assignment because we find sequencing efforts. As a result, research groups have targeted dif-
Pachystylus resolved in a highly supported clade together with ferent parts of the organellar and nuclear genomes, rather than a
Tarenna. common set of loci that would allow integration across data sets.
Our results demonstrate that the standardized set of genes targeted
by the Angiosperms353 probes are highly effective in phylogenetic
Subfamily classification
inference of Gentianales, opening doors for further research on the
Despite the recent advocacy of a three-subfamily division for the evolution of the order. This approach is a cost-effective way to tackle
Rubiaceae (Bremer et al., 1999; Bremer and Eriksson, 2009; Davis large-scale phylogenomic challenges in a coordinated way, as a step
et al., 2009, Rydin et al., 2009; Kainulainen et al., 2013; Mouly towards a complete, integrative, species-level tree of life.
et al., 2014), recent studies (Wikström et al., 2020) and the work
presented here do not support a three- subfamily classification.
We note correspondence between our results and the simplified ACKNOWLEDGMENTS
three-subfamily system proposed by Verdcourt (1958, 1976) and
the study of Robbrecht and Manen (2006). Verdcourt (1958, 1976) We thank the providers of samples for this study, the Plant and
proposed three subfamilies, viz. Cinchonoideae, Rubioideae, and Fungal Trees of Life (PAFTOL) project and horticulture teams at
Kew for support, Martin Cheek for providing information on newly APPENDIX S2. Table showing sequences excluded by Treeshrink.
described species of Rubiaceae, and the two anonymous review- Number of sequences excluded per gene are presented for the data
ers for insightful suggestions. Thanks also to Robyn Cowan, Dion set. Note: the genes highlighted have been entirely excluded as they
Devey (supporting laboratory work), Imalka Kahandawala, Rosy contain less than 25% of the total number of taxa.
Woods (DNA bank sampling).
APPENDIX S3. Nuclear concatenation tree showing relationships
in Gentianales, inferred in IQ-TREE. Bootstrap percentages are pre-
sented above branches, the families are shown in large font, and the
AUTHOR CONTRIBUTIONS
tree tips display the genus, species and tribe for each taxon.
A.A., J.J.C., O.M., and W.J.B. conceived the project and coordi- APPENDIX S4. Plastid concatenation tree showing relationships in
nated the study with contributions from K.K., T.L., S.C.K.S., and Gentianales, inferred in IQ-TREE. Bootstrap percentages are pre-
F.F.; J.J.C. was the overall project manager. O.M. and A.R.Z. car- sented above branches, the families are shown in large font and the
ried out the analyses. J.J.C. and O.M. wrote the methods section. tree tips display the genus, species, and tribe for each taxon.
The text related to particular families was led by T.L., S.C.K.S.,
APPENDIX S5. Plastid coalescent tree showing relationships in
and D.G. for Apocynaceae; K.K., A.P.D., and C.P. for Rubiaceae;
Gentianales, inferred in ASTRAL III. Local posterior probability
L.S. for Gelsemiaceae, Gentianaceae, and Loganiaceae; all of the
values are presented below branches, the families are shown in large
authors contributed to general topics in the introduction and dis-
font, and the tree tips display the genus, species and tribe for each
cussion. A.A., J.J.C., W.J.B., F.F., A.P.D., T.L., S.C.K.S., L.S., K.K.,
taxon. The pie charts show quartet support values for each node
L.P., G.E.B., and A.R.Z. edited the final manuscript. O.M., G.E.B.,
(blue = species tree topology quartet support; red = first alterna-
and N.E. undertook the specimen sampling. Laboratory work was
tive topology quartet support; gray = second alternative topology
carried out by G.E.B., N.E., and L.P. F.F. and W.J.B. supervised the
quartet support). Red stars next to pie charts indicate that the null
PAFTOL project.
hypothesis of a polytomy (equal support for species tree, first alter-
native, and second alternative topology) could not be rejected with
the polytomy test (P > 0.05).
FUNDING INFORMATION
APPENDIX S6. Tanglegram showing a comparison between nu-
This work was funded by grants from the Calleva Foundation clear coalescence tree and nuclear concatenation tree. (A) Nuclear
and the Sackler Trust to the Plant and Fungal Tree of Life Project phylogeny inferred using ASTRAL III. (B) Nuclear phylogeny in-
(PAFTOL) at the Royal Botanic Gardens, Kew, with additional ferred using IQ-TREE. Node support below 100% is indicated by
funding from the Garfield Weston Foundation. A.A. acknowledges color: green (BS ≥ 95%), yellow (75% ≥ BS >95%), red (50% ≥ BS
financial support from the Swedish Research Council (2019-05191), >75%) and black (BS <50%). Families are indicated by color of links:
the Swedish Foundation for Strategic Research (FFL15-0196), and orange (Rubiaceae), purple (Apocynaceae), green (Loganiaceae),
the Royal Botanic Gardens, Kew. L.S. acknowledges funding from red (Gentianaceae), blue (Gelsemiaceae), and gray for outgroups.
the USDA National Institute of Food and Agriculture Hatch proj- See Materials and Methods for the parameters used.
ect (NJ17112) and National Science Foundation (DEB 0317612).
T.L. and S.C.K.S. acknowledge funding from the National Science
Foundation (DEB 1655553/1655223). LITERATURE CITED
Adebowale, A., J. Lamb, A. Nicholas, and Y. Naidoo. 2016. Molecular systematics

of southern African monkey orange Strychnos L. (Loganiaceae). Kew Bulletin
DATA AVAILABILITY
71: 1–16.
Andersson, L., and A. Antonelli. 2005. Phylogeny of the tribe Cinchoneae
All sequences generated for this study are available in the European (Rubiaceae), its position in Cinchonoideae, and description of a new genus,
Nucleotide Archive (ENA) in a custom project area at: https://www. Ciliosemina. Taxon 54: 17–28.
ebi.ac.uk/ena/browser/view/PRJEB35285 and their accession num- Angiosperm Phylogeny Group. 1998. An ordinal classification for the families
bers are provided in Appendix S1. The data set is also available via of flowering plants. Annals of the Missouri Botanical Garden 85: 531–553.
the Kew Tree of Life Explorer (https://treeoflife.kew.org/) using the APG IV, Chase, M. W., M. J. M. Christenhusz, M. F. Fay, J. W. Byng, W. S. Judd, D.
FTP tab (http://sftp.kew.org/pub/paftol/current_release/). Recovery E. Soltis, D. J. Mabberley, et al. 2016. An update of the Angiosperm Phylogeny
statistics were generated using custom scripts available at https:// Group classification for the orders and families of flowering plants: APG IV.
github.com/sidonieB/. Botanical Journal of the Linnean Society 181: 1–20.
Backlund, M., B. Oxelman, and B. Bremer. 2000. Phylogenetic relationships
within the Gentianales based on ndhF and rbcL sequences, with particular
reference to Loganiaceae. American Journal of Botany 87: 1029–1043.
SUPPORTING INFORMATION Baker, W. J., P. Bailey, V. Barber, A. Barker, S. Bellot, D. Bishop, L. R. Botigué, et al.
2021. A comprehensive phylogenomic platform for exploring the angio-
Additional Supporting Information may be found online in the sperm tree of life. Systematic Biology syab035; https://doi.org/10.1093/sysbi
supporting information tab for this article. o/syab035 [in press].
Bankevich, A., S. Nurk, D. Antipov, A. A. Gurevich, M. Dvorkin, A. S. Kulikov,
APPENDIX S1. Table showing sequencing statistics. Number of V. M. Lesin, et al. 2012. SPAdes: a new genome assembly algorithm and its
reads, read depth, contigs per taxon, and other metrics are shown applications to single-cell sequencing. Journal of Computational Biology 19:
for the Gentianales dataset. 455–477.
Bolger, A. M., M. Lohse, and B. Usadel. 2014. Trimmomatic: a flexible trimmer for Doyle, J. J., and J. L. Doyle. 1987. A rapid DNA isolation procedure for small
Illumina sequence data. Bioinformatics 30: 2114–2120. quantities of fresh leaf tissue. Phytochemistry Bulletin 19: 11–15.
Bremer, B., and T. Eriksson. 2009. Time tree of Rubiaceae: phylogeny and dat- Endress, M. E., U. Meve, D. J. Middleton, and S. Liede- Schumann. 2018.
ing the family, subfamilies, and tribes. International Journal of Plant Sciences Apocynaceae. In J. Kadereit and V. Bittrich [eds.], Flowering plants. Eudicots.
170: 766–793. Apiales, Gentianales (except Rubiaceae), vol. 15, The families and gen-
Bremer, B., R. K. Jansen, B. Oxelman, M. Backlund, H. Lantz, and K. J. Kim. 1999. era of vascular plants, 207-411. Springer International Publishing, Cham,
More characters or more taxa for a robust phylogeny—case study from the Switzerland.
coffee family (Rubiaceae). Systematic Biology 48: 413–435. Fishbein, M., T. Livshultz, S. C. K. Straub, A. O. Simões, J. Boutte, A. McDonnell,
Brewer, G. E., J. J. Clarkson, O. Maurin, A. R. Zuntini, V. Barber, S. Bellot, N. Biggs, and A. Foote. 2018. Evolution on the backbone: Apocynaceae phylogenomics
et al. 2019. Factors affecting targeted sequencing of 353 nuclear genes from and new perspectives on growth forms, flowers, and fruits. American Journal
herbarium specimens spanning the diversity of angiosperms. Frontiers in of Botany 105: 495–513.
Plant Science 10: 1102. Franke, J., J. Kim, J. P. Hamilton, D. Zhao, G. M. Pham, K. Wiegert-Rininger, E.
Bridson, D. M., and E. Robbrecht. 1985. Further notes on the tribe Pavetteae Crisovan, et al. 2019. Gene discovery in gelsemium highlights conserved
(Rubiaceae). Bulletin du Jardin botanique national de Belgique/Bulletin van gene clusters in monoterpene indole alkaloid biosynthesis. Chembiochem 20:
de Nationale Plantentuin van België 55: 83–115. 83–87.
Brown, R. W. 1929. Additions to the flora of the Green River formation. U.S. Frasier, C. L. 2008. Evolution and systematics of the angiosperm order Gentianales
Geological Survey Professional Paper 154: 279–292. with an in-depth focus on Loganiaceae and its species-rich and toxic genus
Brown, R. W. 1934. The recognizable species of the Green River Flora. U.S. Strychnos. Ph.D. dissertation, Rutgers University, New Brunswick, NJ, USA.
Geological Survey Professional Paper 185–C: 45–77. Frodin, D. 2004. History and concepts of big plant genera. Taxon 53: 741–752.
Bruyns, P. V. 1999. The systematic position of Eustegia R.Br. (Apocynaceae - Gibbons, K. L., M. J. Henwood, and B. J. Conn. 2012. Phylogenetic relationships
Asclepiadoideae). Botanische Jahrbücher für Systematik, Pflanzengeschichte in Loganieae (Loganiaceae) inferred from nuclear ribosomal and chloroplast
und Pflanzengeographie. 121: 19–44. DNA sequence data. Australian Systematic Botany 25: 331–340.
Buerki, S., M. W. Callmander, P. Acevedo-Rodriguez, P. P. Lowry, J. Munzinger, Hoang, D. T., O. Chernomor, A. von Haeseler, B. Q. Minh, and L. S. Vinh. 2018.
P. Bailey, O. Maurin, et al. 2021. An updated infra-familial classification of UFBoot2: improving the ultrafast bootstrap approximation. Molecular
Sapindaceae based on targeted enrichment data. American Journal of Botany Biology and Evolution 35: 518–522.
108: 1–18. https://doi.org/10.1002/ajb2.1693. Jansen, M. E. 1979. A revision of the genus Kajewskiella (Rubiaceae). Blumea 25:
Camacho, C., G. Coulouris, V. Avagyan, N. Ma, J. Papadopoulos, K. Bealer, 283–294.
and T. L. Madden. 2009. BLAST+: architecture and applications. BMC Jiang, X., S. V. Edwards, and L. Liu. 2020. The multispecies coalescent model out-
Bioinformatics 10: 421. performs concatenation across diverse phylogenomic data sets. Systematic
Canales, N. A., T. N. G. Hansen, C. Cornett, K. Walker, F. Driver, A. Antonelli, Biology 69: 795–812.
C. Maldonado, et al. 2020. Historical chemical annotations of Cinchona Johnson, M. G., E. M. Gardner, Y. Liu, R. Medina, B. Goffinet, A. J. Shaw, N. J.
bark collections are comparable to results from current day high-pressure Zerega, and N. J. Wickett. 2016. HybPiper: Extracting coding sequence and
liquid chromatography technologies. Journal of Ethnopharmacology 249: introns for phylogenetics from high-throughput sequencing reads using tar-
112375. get enrichment. Applications in Plant Science 4: 1600016.
Chase, M. W., and H. G. Hills. 1991. Silica gel: an ideal desiccant for preserving Johnson, M. G., L. Pokorny, S. Dodsworth, L. R. Botigué, R. S. Cowan, A. Devault,
field-collected leaves for use in molecular studies. Taxon 41: 215–220. W. L. Eiserhardt, et al. 2019. A universal probe set for targeted sequencing
Chase, M. W., D. E. Soltis, R. G. Olmstead, D. Morgan, D. H. Les, B. D. Mishler, of 353 nuclear genes from any flowering plant designed using k-medoids
M. R. Duvall, et al. 1993. Phylogenetics of seed plants: an analysis of nucleo- clustering. Systematic Biology 68: 594–606.
tide sequences from the plastid gene rbcL. Annals of the Missouri Botanical Junier, T., and E. M. Zdobnov. 2010. The Newick utilities: high-throughput
Garden 80: 528–580. phylogenetic tree processing in the UNIX shell. Bioinformatics 26:
Cheek, M., E. Nic Lughadha, P. Kirk, H. Lindon, J. Carretero, B. Looney, B. 1669–1670.
Douglas, et al. 2020. New scientific discoveries: plants and fungi. Plants, Kainulainen, K., S. G. Razafimandimbison, and B. Bremer. 2013. Phylogenetic re-
People, Planet 2: 371–388. lationships and new tribal delimitations in subfamily Ixoroideae (Rubiaceae).
Chernomor, O., A. von Haeseler, and B. Q. Minh. 2016. Terrace aware data struc- Botanical Journal of the Linnean Society 173: 387–406.
ture for phylogenomic inference from supermatrices. Systematic Biology 65: Leebens- Mack, J. H., M. S. Barker, E. J. Carpenter, M. K. Deyholos, M. A.
997–1008. Gitzendanner, S. W. Graham, I. Grosse, et al. 2019. One thousand plant tran-
Clarkson, J. J., A. R. Zuntini, O. Maurin, S. R. Downie, G. M. Plunkett, A. A. scriptomes and the phylogenomics of green plants. Nature 574: 679–685.
Nicolas, J. F. Smith, et al. 2021. A higher-level nuclear phylogenomic study Leeuwenberg, A. J. M., and P. W. Leenhouts. 1980. Taxonomy. In A. J. M.
of the carrot family (Apiaceae). American Journal of Botany 108 (in press). Leeuwenberg [ed.], Engler and Prantl’s Die natürlichen Pflanzenfamilien, ed.
Cragg, G. M., and D. J. Newman. 2005. Plants as a source of anti-cancer agents. 2, 28b(1), 8–96. Duncker & Humblot, Berlin, Germany.
Journal of Ethnopharmacology 100: 72–79. Li, H.-T., T.-S. Yi, L.-M. Gao, P.-F. Ma, T. Zhang, J.-B. Yang, M. A. Gitzendanner,
Davis, A. P., and D. M. Bridson. 2007. Rubiaceae. In V. H. Heywood, R. K. et al. 2019. Origin of angiosperms and the puzzle of the Jurassic gap. Nature
Brummitt, A. Culham, and O. Seberg [eds.], Flowering plants of the world, Plants 5: 461–470.
284–286. Royal Botanic Gardens, Kew, UK. Liede, S. 2001. Subtribe Astephaninae (Apocynaceae- Asclepiadoideae) re-
Davis, A. P., R. Govaerts, D. M. Bridson, M. Ruhsam, J. Moat, and N. A. Brummitt. considered: new evidence based on cpDNA spacers. Annals of the Missouri
2009. A global assessment of distribution, diversity, endemism, and tax- Botanical Garden 88: 657–668.
onomic effort in the Rubiaceae. Annals of Missouri Botanical Garden 96: Liede, S., and A. Täuber. 2002. Circumscription of the genus Cynanchum
68–78. (Apocynaceae-Asclepiadoideae). Systematic Botany 27: 789–800.
De Block, P., S. G. Razafimandimbison, S. Janssens, H. Ochoterena, E. Robbrecht, Liede-Schumann, S., M. Nikolaus, U. C. S. e Silva, A. Rapini, R. D. Mangelsdorff,
and B. Bremer. 2015. Molecular phylogenetics and generic assessment in the and U. Meve. 2014. Phylogenetics and biogeography of the genus Metastelma
tribe Pavetteae (Rubiaceae). Taxon 64: 79–95. (Apocynaceae-Asclepiadoideae-Asclepiadeae: Metastelmatinae). Systematic
Del Rio, C., T.-X. Wang, J. Liu, S.-Q. Liang, R. A. Spicer, F.-X. Wu, Z.-K. Zhou, Botany 39: 594–612.
and T. Su. 2020. Asclepiadospermum gen. nov., the earliest fossil record of Liede-Schumann, S., A. Rapini, D. J. Goyder, and M. W. Chase. 2005. Phylogenetics
Asclepiadoideae (Apocynaceae) from the early Eocene of central Qinghai- of the new world subtribes of Asclepiadeae (Apocynaceae-Asclepiadoideae):
Tibetan Plateau, and its biogeographic implications. American Journal of Metastelmatinae, Oxypetalinae, and Gonolobinae. Systematic Botany 30:
Botany 107: 126–138. 184–195.
Livshultz, T. 2010. The phylogenetic position of milkweeds (Apocynaceae sub- Puff, C., A. Igersheim, R. Buchner, and U. Rohrhofer. 1995. The united stamens of
families Secamonoideae and Asclepiadoideae): Evidence from the nucleus Rubiaceae. Morphology, anatomy; their role in pollination ecology. Annals of
and chloroplast. Taxon 59: 1016–1030. the Missouri Botanical Garden 82: 357–382.
Livshultz, T., J. V. Mead, D. J. Goyder, and M. Brannin. 2011. Climate niches of Qiu, Y.-L., L. Li, B. Wang, J.-Y. Xue, T. A. Hendry, R.-Q. Li, J. W. Brown, et al. 2010.
milkweeds with plesiomorphic traits (Secamonoideae; Apocynaceae) and Angiosperm phylogeny inferred from sequences of four mitochondrial
the milkweed sister group link ancient African climates and floral evolution. genes. Journal of Systematics and Evolution 48: 391–425.
American Journal of Botany 98: 1966–1977. Ramírez-Barahona, S., H. Sauquet, and S. Magallón. 2020. The delayed and geo-
Livshultz, T., D. J. Middleton, M. E. Endress, and J. K. Williams. 2007. Phylogeny graphically heterogeneous diversification of flowering plant families. Nature
of Apocynoideae and the APSA clade (Apocynaceae s.l.). Annals of the Ecology & Evolution 4: 1232–1238.
Missouri Botanical Garden 94: 324–359. Rapini, A., M. W. Chase, D. J. Goyder, and J. Griffiths. 2003. Asclepiadeae
Manns, U., and B. Bremer. 2010. Towards a better understanding of intertribal classification: evaluating the phylogenetic relationships of New World
relationships and stable tribal delimitations within Cinchonoideae ss Asclepiadoideae (Apocynaceae). Taxon 52: 33–50.
(Rubiaceae). Molecular Phylogenetics and Evolution 56: 21–39. Rapini, A., C. van den Berg, and S. Liede-Schumann. 2007. Diversification of
Mark, J., A. C. Newton, S. Oldfield, and M. Rivers. 2014. The international tim- Asclepiadoideae (Apocynaceae) in the New World. Annals of the Missouri
ber trade: a working list of commercial timber tree species. Botanic Gardens Botanical Garden 94: 407–422.
Conservation International, Richmond, UK. Razafimandimbison, S. G., K. Kainulainen, K. M. Wong, K. Beaver, and B. Bremer.
Mai, U., and S. Mirarab. 2018. TreeShrink: fast and accurate detection of outlier 2011. Molecular support for a basal grade of morphologically distinct, mono-
long branches in collections of phylogenetic trees. BMC Genomics 19: 272. typic genera in the species-rich Vanguerieae alliance (Rubiaceae, Ixoroideae):
Maurin, O., A. Anest, S. Bellot, E. Biffin, G. Brewer, T. Charles-Dominique, R. Its systematic and conservation implications. Taxon 60: 941–952.
S. Cowan, et al. 2021. A nuclear phylogenomic study of the angiosperm Razafimandimbison, S. G., and C. Rydin. 2019. Molecular based assessments
order Myrtales, exploring the potential and limitations of the universal of tribal and generic limits and relationships in Rubiaceae (Gentianales):
Angiosperms353 probe set. American Journal of Botany 108 (in press). polyphyly of Pomazoteae and paraphyly of Ophiorrhizeae and Ophiorrhiza.
McKey, D. 1980. Origins of novel alkaloid types: a mechanism for rapid phenotypic Taxon 68: 72–91.
evolution of plant secondary compounds. American Naturalist 115: 754–759. Razafimandimbison, S. G., C. M. Taylor, N. Wikström, T. Pailler, A. Khodabandeh,
Merckx, V. S., J. Kissling, H. Hentrich, S. B. Janssens, C. B. Mennes, C. D. Specht, and B. Bremer. 2014. Phylogeny and generic limits in the sister tribes
and E. F. Smets. 2013. Phylogenetic relationships of the mycoheterotro- Psychotrieae and Palicoureeae (Rubiaceae): Evolution of schizocarps in
phic genus Voyria and the implications for the biogeographic history of Psychotria and origins of bacterial leaf nodules of the Malagasy species.
Gentianaceae. American Journal of Botany 100: 712–721. American Journal of Botany 101: 1102–1126.
Merrill, E. D., and L. M. Perry. 1947. Kajewskiella, a new rubiaceous genus from Refulio-Rodriguez, N. F., and R. G. Olmstead. 2014. Phylogeny of Lamiidae.
the Solomon Islands. Journal of the Arnold Arboretum 28: 331–332. American Journal of Botany 101: 287–999.
Meve, U., and S. Liede. 2004. Subtribal division of Ceropegieae (Apocynaceae- Ridsdale, C. E. 1975. Notes on New Guinea Rubiaceae: Chaetostachydium.
Asclepiadoideae). Taxon 53: 61–72. Blumea 22: 267–269.
Minh, B. Q., H. A. Schmidt, O. Chernomor, D. Schrempf, M. D. Woodhams, A. Ridsdale, C. E., R. C. B. van den Brink, and J. Koek-Noorman. 1972. Notes on New
Von Haeseler, and R. Lanfear. 2020. IQ-TREE 2: New models and efficient Guinea Rubiaceae: Versteegia and Maschalodesme. Blumea 20: 339–348.
methods for phylogenetic inference in the genomic era. Molecular Biology Robbrecht, E. 1988. Tropical woody Rubiaceae. Opera Botanica Belgica 1: 1–271.
and Evolution 37: 1530–1534. Robbrecht, E., and J. F. Manen. 2006. The major evolutionary lineages of the coffee
Mirarab, S. 2019. Species tree estimation using ASTRAL: practical consider- family (Rubiaceae, angiosperms). Combined analysis (nDNA and cpDNA)
ations. arXiv arXiv:1904.03826 [Preprint]. to infer the position of Coptosapelta and Luculia, and supertree construction
Mirarab, S., and T. Warnow. 2015. ASTRAL-II: coalescent-based species tree esti- based on rbcL, rps16, trnL-trnF and atpB-rbcL data. A new classification in
mation with many hundreds of taxa and thousands of genes. Bioinformatics two subfamilies, Cinchonoideae and Rubioideae. Systematics and Geography
31: i44–52. of Plants 76: 85–145.
Molina, J., and L. Struwe. 2009. Utility of secondary structure in phylogenetic Robbrecht, E., and C. Puff. 1986. Survey of the Gardenieae and related tribes
reconstructions using nrDNA ITS sequences—an example from Potalieae (Rubiaceae). Botanische Jahrbücher fur Systematik, Pflanzengeschichte und
(Gentianaceae: Asteridae). Systematic Botany 34: 414–428. Pflanzengeographie 108: 63–137.
Mouly, A., K. Kainulainen, C. Persson, A. P. Davis, K. M. Wong, S. G. Rokas, A., and S. B. Carroll. 2006. Bushes in the tree of life. PLoS Biology 4:
Razafimandimbison, and B. Bremer. 2014. Phylogenetic structure and clade 1899–1904.
circumscriptions in the Gardenieae complex (Rubiaceae). Taxon 63: 801–818. Rova, J. H., P. G. Delprete, L. Andersson, and V. A. Albert. 2002. A trnL-F cpDNA
Murphy, B., F. Forest, T. Barraclough, J. Rosindell, S. Bellot, R. Cowan, et al. sequence study of the Condamineeae-Rondeletieae-Sipaneeae complex with
2019. A phylogenomic analysis of Nepenthes (Nepenthaceae). Molecular implications on the phylogeny of the Rubiaceae. American Journal of Botany
Phylogenetics and Evolution 144: 1066682. 89: 145–159.
Oxelman, B., and B. Bremer. 2000. Discovery of paralogous nuclear gene se- Rydin, C., K. Kainulainen, S. G. Razafimandimbison, J. E. Smedmark, and B.
quences coding for the second-largest subunit of RNA polymerase II (RPB2) Bremer. 2009. Deep divergences in the coffee family and the systematic posi-
and their phylogenetic utility in Gentianales of the asterids. Molecular tion of Acranthera. Plant Systematics and Evolution 278: 101–123.
Biology and Evolution 17: 1131–1145. Rydin, C., N. Wikström, and B. Bremer. 2017. Conflicting results from mito-
Paradis, E., and K. Schliep. 2019. ape 5.0: an environment for modern phyloge- chondrial genomic data challenge current views of Rubiaceae phylogeny.
netics and evolutionary analyses in R. Bioinformatics 35: 526–528. American Journal of Botany 104: 1522–1532.
Patel, K., D. Laloo, G. K. Singh, M. Gadewar, and D. Kumar. 2017. A review on Sayyari, E., and S. Mirarab. 2018. Testing for polytomies in phylogenetic species
medicinal uses, analytical techniques and pharmacological activities of trees using quartet frequencies. Genes 9: 132.
Strychnos nux-vomica Linn.: a concise report. Chinese Journal of Integrative Shah, T., J. V. Schneider, G. Zizka, O. Maurin, W. Baker, F. Forest, G. E. Brewer,
Medicine. https://doi.org/10.1007/s11655-016-2514-1. et al. 2021. Joining forces in Ochnaceae phylogenomics: a tale of two targeted
Pérez-Escobar, O. A., S. Dodsworth, D. Bogarín, S. Bellot, J. A. Balbuea, R. J. Schley, sequencing probe kits. American Journal of Botany 108: 1–15. https://doi.
I. A. Kikuchi, et al. 2021. Hundreds of nuclear and plastid loci yield novel in- org/10.1002/ajb2.1682.
sights into orchid relationships. American Journal of Botany 108 (in press). Simões, A. O., T. Livshultz, E. Conti, and M. E. Endress. 2007. Phylogeny and
POWO. 2020. Plants of the World Online. Facilitated by the Royal Botanic systematics of the Rauvolfioideae (Apocynaceae) based on molecular and
Gardens, Kew. Website: http://www.plantsoftheworldonline.org/ [accessed morphological evidence. Annals of the Missouri Botanical Garden 94:
22 September 2020]. 268–297.
Smith, M. E. 2003. 40Ar/39 Ar geochronology of the Eocene Green River Team, R.C., 2020. R: A Language and Environment for Statistical Computing.
Formation, Wyoming. Geological Society of America Bulletin 115: 549–565. R Foundation for Statistical Computing, Vienna, Austria: Website: https://
Soltis, D. E., S. A. Smith, N. Cellinese, K. J. Wurdack, D. C. Tank, S. F. Brockington, www.R-project.org.
N. F. Refulio-Rodriguez, et al. 2011. Angiosperm phylogeny: 17 genes, 640 van der Heijden, R., D. I. Jacobs, W. Snoeijer, D. Hallard, and R. Verpoorte. 2004.
taxa. American Journal of Botany 98: 704–730. The Catharanthus alkaloids: pharmacognosy and biotechnology. Current
Slater, G. S., and E. Birney. 2005. Automated generation of heuristics for biologi- Medicinal Chemistry 11: 607–628.
cal sequence comparison. BMC Bioinformatics 6: 31. Vegro, C. L. R., and L. Florêncio de Almeida. 2020. Global coffee market: socio-
Straub, S. C. K., R. C. Cronn, C. Edwards, M. Fishbein, and A. Liston. 2013. economic and cultural dynamics. In L. Florêncio de Almeida and E. E. Spers
Horizontal transfer of DNA from the mitochondrial to the plastid genome [eds.], Coffee consumption and industry strategies in Brazil, 3–19. Woodhead
and its subsequent evolution in milkweeds (Apocynaceae). Genome Biology Publishing, Sawston, UK.
and Evolution 5: 1872–1885. Verdcourt, B. 1958. Remarks on the classification of the Rubiaceae. Bulletin du
Straub, S. C. K., M. J. Moore, P. S. Soltis, D. E. Soltis, A. Liston, and T. Livshultz. Jardin Botanique de l’État à Bruxelles 6: 173–196.
2014. Phylogenetic signal detection from an ancient rapid radiation: effects Verdcourt, B. 1976. Notes on African Rubiaceae. Kew Bulletin 31: 181–186.
of noise reduction, long-branch attraction, and model selection in crown Verdcourt, B. 1989. Rubiaceae (Rubioideae). In: Launert E. (ed.) Flora Zambesiaca
clade Apocynaceae. Molecular Phylogenetics and Evolution 80: 169–185. 5(1). Flora Zambesiaca managing committee: London.
Struwe, L.. 2014. Classification and evolution of the Gentianaceae. In J. J. von Hagen, K. B., and J. W. Kadereit. 2003. The diversification of Halenia
Rybczyński, M. R. Davey, and A. Mikula [eds.], The Gentianaceae: charac- (Gentianaceae): Ecological opportunity versus key innovation. Evolution 57:
terization and ecology, vol. 1, 13–35. Springer Verlag, Heidelberg, Germany. 2507–2518.
Struwe, L. 2018. Gelsemiaceae. In K. Kubitzki [ed.], Families and genera of flow- Wang, L. G., T. T. Y. Lam, S. Xu, Z. Dai, L. Zhou, T. Feng, P. Guo, et al. 2020. treeio:
ering plants; W. Kadereit and V. Bittrich [eds.], vol. XV, Asteridae, Gentianales an R package for phylogenetic tree input and output with richly annotated
(except Rubiales), 447–452. Springer Verlag, Berlin, Germany. and associated data. Molecular Biology and Evolution 37: 599–603.
Struwe, L., V. A. Albert, and B. Bremer. 1994. Cladistics and family level classifica- Wendel, J. F., and J. J. Doyle. 1998. Phylogenetic incongruence: window into ge-
tion of the Gentianales. Cladistics 10: 175–206. nome history and molecular evolution. In D. E. Soltis, P. S. Soltis, and J. J.
Struwe, L., V. A. Albert, M. F. Calió, C. Frasier, K. B. Lepis, K. G. Mathews, and J. Doyle [eds.], Molecular systematics of plants II: DNA sequencing, 265–296.
R. Grant. 2009. Evolutionary patterns in neotropical Helieae (Gentianaceae): Kluwer, Boston, MA, USA.
evidence from morphology, chloroplast and nuclear DNA sequences. Taxon Wikström, N., B. Bremer, and C. Rydin. 2020. Conflicting phylogenetic signals
58: 479–499. in genomic data of the coffee family (Rubiaceae). Journal of Systematics and
Struwe, L., K. Gibbons, B. Conn, and T. Motley. 2018. Loganiaceae (includ- Evolution 58: 440–460.
ing Antoniaceae, Geniostomaceae, Spigeliaceae, and Strychnaceae). In K. Wikström, N., K. Kainulainen, S. G. Razafimandimbison, J. E. Smedmark, and
Kubitzki [ed.], Families and genera of flowering plants; W. Kadereit and V. B. Bremer. 2015. A revised time tree of the asterids: establishing a temporal
Bittrich [eds.], vol. XV, Asteridae, Gentianales (except Rubiales), 511–526. framework for evolutionary studies of the coffee family (Rubiaceae). PLoS
Springer Verlag, Berlin, Germany. One 10: e0126690.
Struwe, L., J. W. Kadereit, J. Klackenberg, S. Nilsson, M. Thiv, K. B. von Hagen, Yang, L.-L., H.-L. Li, L. Wei, T. Yang, D.-Y. Kuang, M.-H. Li, Y.-Y. Liao, et al. 2016. A
and V. A. Albert. 2002. Systematics, character evolution, and biogeography of supermatrix approach provides a comprehensive genus-level phylogeny for
Gentianaceae, including a new tribal and subtribal classification. In L. Struwe Gentianales. Journal of Systematics and Evolution 54: 400–415.
and V. A. Albert [eds.], Gentianaceae: systematics and natural history, 21– Yu, G., 2019. ggimage: Use Image in ‘ggplot2’ R package version 0.2.1.
309. Cambridge University Press, Cambridge, UK. Yu, G., D. K. Smith, H. Zhu, Y. Guan, and T. T. Y. Lam. 2017. ggtree: an R
Struwe, L., and J. Pringle. 2018. Gentianaceae. In K. Kubitzki [ed.], Families package for visualization and annotation of phylogenetic trees with their
and genera of flowering plants; W. Kadereit and V. Bittrich [eds.], vol. XV, covariates and other associated data. Methods in Ecology and Evolution
Asteridae, Gentianales (except Rubiales), 453–503. Springer Verlag, Berlin. 8: 28–36.
Stull, G. W., R. Duno de Stefano, D. E. Soltis, and P. S. Soltis. 2015. Resolving basal Zhang, C., M. Rabiee, E. Sayyari, and S. Mirarab. 2018. ASTRAL-III: polynomial
lamiid phylogeny and the circumscription of Icacinaceae with a plastome- time species tree reconstruction from partially resolved gene trees. BMC
scale data set. American Journal of Botany 102: 1794–1813. Bioinformatics 19: 153.
Stull, G. W., P. S. Soltis, D. E. Soltis, M. A. Gitzendanner, and S. A. Smith. 2020. Zhang, C., T. Zhang, F. Luebert, Y. Xiang, C.-H. Huang, Y. Hu, M. Rees, et al.
Nuclear phylogenomic analyses of asterids conflict with plastome trees and 2020. Asterid phylogenomics/phylotranscriptomics uncover morpholog-
support novel relationships among major lineages. American Journal of ical evolutionary histories and support phylogenetic placement for nu-
Botany 107: 790–805. merous whole-genome duplications. Molecular Biology and Evolution 37:
Surveswaran, S., M. Sun, G. W. Grimm, and S. Liede-Schumann. 2014. On the 3188–3210.
systematic position of some Asian enigmatic genera of Asclepiadoideae Zuntini, A. R., L. P. Frankel, L. Pokorny, F. Forest, and W. J. Baker. 2021. A com-
(Apocynaceae). Botanical Journal of the Linnean Society 174: 601–619. prehensive phylogenomic study of the monocot order Commelinales with a
Tange, C. 1995. The identity of Siderobombyx and a new species of Xanthophytum new classification of Commelinaceae. American Journal of Botany 108 (in
(Rubiaceae). Nordic Journal of Botany 15: 575–581. press). https://doi.org/10.1002/ajb2.1698.
View publication stats

Antonellietal2021 (1)

Uploaded by

Copyright:

Available Formats

Antonellietal2021 (1)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Antonellietal2021 (1)

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Settling a family feud: a high-level phylogenomic framework for the

Article in American Journal of Botany · July 2021

Alexandre Antonelli James Clarkson

SEE PROFILE SEE PROFILE

Kent Kainulainen Olivier Maurin

SEE PROFILE SEE PROFILE

The user has requested enhancement of the downloaded file.

INVITED SPECIAL ARTICLE

Settling a family feud: a high-­level phylogenomic framework

doi:10.1002/ajb2.1697 target sequence capture; tree of life.

FIGURE 2. Alternative topologies for four of the five Gentianales fami-

TABLE 1. Voucher specimens associated with sampling.

Percentage of target length

Samples Any length 25% or longer 50% or longer 75% or longer

Apocynaceae Gelsemiaceae Gentianaceae Loganiaceae Rubiaceae Outgroup

A Coptosapelta diffusa Tarenna pulchra B

Apocynaceae Gentianaceae Rubiaceae 100% ● >=75% <95% ● <50%

Adebowale, A., J. Lamb, A. Nicholas, and Y. Naidoo. 2016. Molecular systematics

View publication stats

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Settling a family feud: a high-level phylogenomic framework