Content uploaded by Quentin Cronk
Author content
All content in this area was uploaded by Quentin Cronk on Aug 03, 2023
Content may be subject to copyright.
Accepted Manuscript
© The Author(s) 2023. Published by Oxford University Press, on behalf of the Society of
Systematic Biologists. All rights reserved. For permissions, please email:
journals.permissions@oup.com
Phylogenomics reveals patterns ancient hybridization and differential diversification
that contribute to phylogenetic conflict in willows, poplars, and close relatives.
Brian J. Sanderson*1,8,11, Diksha Ghambir1, Guanqiao Feng1, Nan Hu1, Quentin C. Cronk2,
Diana M. Percy2, Francisco Molina Freaner3, Matthew G. Johnson1, Lawrence B. Smart4,
Ken Keefover-Ring5, Tongming Yin6, Tao Ma7, Stephen P. DiFazio8, Jianquan Liu7,9,
Matthew S. Olson*1,10
1Department of Biological Sciences, Texas Tech University, Lubbock, TX 79409-3131 USA
2Department of Botany, University of British Columbia, Vancouver, BC, V6T 1Z4 Canada
3Universidad Nacional Automoa de Mexico, Hermosilla, Mexico
4Horticulture Section, School of Integrative Plant Science, Cornell University, Cornell
AgriTech, Geneva, New York 14456 USA
5Departments of Botany and Geography, University of Wisconsin-Madison, Madison, WI
53706, USA
6Key Laboratory of Tree Genetics and Biotechnology of Jiangsu Province and Education
Department of China, Nanjing Forestry University, Nanjing, China
7Key Laboratory of Bio-Resource and Eco-Environment of Ministry of Education & College
of Life Sciences, Sichuan University, Chengdu 610065, China
8Department of Biology, West Virginia University, Morgantown, WV, 26506 USA
9State Key Laboratory of Grassland Agro-Ecosystem, Institute of Innovation Ecology &
College of Life Sciences, Lanzhou University, Lanzhou 730000, China
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
2
10Author for correspondence: matt.olson@ttu.edu
*These authors contributed equally to this work.
11Current The Jackson Laboratory for Genomic Medicine, 10 Discovery Drive, Farmington,
CT 06032 USA
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
3
Abstract
Despite the economic, ecological, and scientific importance of the genera Salix L.
(willows) and Populus L. (poplars, cottonwoods, and aspens) Salicaceae, we know little
about the sources of differences in species diversity between the genera and of the
phylogenetic conflict that often confounds estimating phylogenetic trees. Salix subgenera and
sections, in particular, have been difficult to classify, with one recent attempt termed a
Vetrix and Chamaetia.
Here we use targeted sequence capture to understand the evolutionary history of this portion
of the Salicaceae plant family. Our phylogenetic hypothesis was based on 787 gene regions
and identified extensive phylogenetic conflict among genes. Our analysis supported some
previously described subgeneric relationships and confirmed polyphyly of others. Using an
fbranch analysis we identified several cases of hybridization in deep branches of the phylogeny,
which likely contributed to discordance among gene trees. In addition, we identified a rapid
increase in diversification rate near the origination of the Vetrix-Chamaetia clade in Salix.
This region of the tree coincided with several nodes that lacked strong statistical support,
indicating a possible increase in incomplete lineage sorting due to rapid diversification. The
extraordinary level of both recent and ancient hybridization in both Salix and Populus have
played important roles in the diversification and diversity in these two genera.
Keywords: Salicaceae, hybridization, f-branch test, sequence capture, ASTRAL species tree,
concatenated tree
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
4
This study is dedicated to the memory of George W. Argus (1929-2022) whose lifelong
pursuit of understanding diversity in Salix laid the foundation for future salicologists.
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
5
Introduction
As methods to assess the congruence among the genealogical histories of genes across
species have matured (Degnan et al. 2009; Young et al. 2020), the curious association
between phylogenetic conflict and rapid diversification has suggested a link between
population genetic and macroevolutionary processes (Parins-Fukuchi et al. 2021). Although
most genomic regions are expected to reflect the speciation and diversification history of a
taxonomic group (species tree), two primary factors contribute to conflict between gene
genealogies and species history. Incomplete lineage sorting (ILS) results from the persistence
of polymorphism across multiple diversification events and the subsequent random fixation
of polymorphism among different lineages (Wu 1991). The influence of ILS is particularly
strong during periods of rapid speciation, when effective population sizes are large and when
long-term balancing selection results in persistence of polymorphisms (Edwards 2009; Pease
et al. 2013; Wang et al. 2020). Second, interspecific gene flow due to hybridization also has
the potential to generate discordance among high proportions of gene trees, (McVay et al.
2017; Morales-Briones et al. 2020). Unlike ILS, however, when hybridization is
accompanied by backcrossing to one parental species, it generates biased frequencies of the
two 4-taxon topologies that are inconsistent with the species tree (e.g. ABBA-BABA;
Hudson 1983; Green et al. 2010; Durand et al. 2011; Patterson et al. 2012). When taxonomic
groups have hybridized throughout diversification, the effects of hybridization on gene tree
conflict can span multiple species within a clade (Malinsky et al. 2018). The impacts of
hybridization may be particularly common in plants, where fertility of interspecific crosses
may be maintained well after speciation (Grant 1981).
Contemporary populations of willows (Salix) and poplars (Populus) are widely known
to hybridize, with important evolutionary and ecological consequences along hybrid zones
(Brunsfeld et al. 1992; Hardig et al. 2000; Schweitzer et al. 2004; Lexer et al. 2010; Chhatre
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
6
et al. 2018; Wang et al. 2020). The impacts of hybridization on evolution and diversification
are evident in both genera, where multiple independent chloroplast capture events have
occurred early in their diversification (Smith et al. 1990; Brunsfeld et al. 1992; Liu et al.
2017; Wang et al. 2020; Gulyaev et al. 2022). This history of hybridization creates challenges
for the reconstruction of phylogenetic histories of willows and poplars (Percy et al. 2014).
Recent progress using genome-wide data sets, however, have constructed well-supported
taxonomic relationships within both genera (Wagner et al. 2020; Wang et al. 2020; Wagner et
al. 2021a; Gulyaev et al. 2022; Wang et al. 2022), but the sources of conflict among gene
trees have not been fully investigated, especially in Salix.
Willows and poplars are integral components of temperate, boreal, and arctic
ecosystems throughout the northern hemisphere and many species have significant cultural,
medical, and economic importance (Stettler et al. 1996; Argus 1997). Phytochemical
diversity in these genera spans an impressive array of secondary metabolites, including
aspirin and its derivatives (Desborough et al. 2017) and defensive chemicals such as phenolic
glycosides, condensed tannins and hydroxycinnamate derivatives (Tsai et al. 2006; Philippe
et al. 2007; Boeckler et al. 2011; Keefover-Ring et al. 2022). Morphological variation in both
genera ranges from dwarf creeping Salix in alpine and arctic zones that were once categorized
as a separate genus (Stettler et al. 1996; Argus 1997) to large Populus trees in subtropical
zones. Many more species are recognized in Salix (approx. 450-520 species; Argus 2010)
than in Populus (approx. 100 species; Shu 1999), suggesting that either Salix began to
diversify much earlier than Populus or speciation rates have increased in Salix.
Salix and Populus are the two largest genera in the Salicaceae and all but one species
across both genera are dioecious (Rohwer et al. 1984). The Salicaceae whole genome
duplication unites all genera except Azara, which lacks the duplication (Cronk et al. 2015).
Within the Salicaceae, Salix and Populus are united by a striking synapomorphy of flowers
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
7
organized into aments or catkins (Meeuse 1975; Argus 2010; Eckenwalder 2010; Cronk et al.
2015) with seed dispersal via wind. This differs from their closest relative Idesia, which
produces fleshy animal-dispersed fruits. Salix and Populus differ in that pollen is dispersed
by insects or by both insects and wind in Salix and only by wind in Populus (Sacchi et al.
1988; Tamura et al. 2000; Karrenberg et al. 2002), suggesting that factors underlying
pollination mode may drive differences in the diversification rate between the two genera
(Friedman et al. 2009; Wessinger 2021). The reduced floral structures in both genera exhibit
relatively low variability and are used to discriminate among species only at the broadest
taxonomic levels (Eckenwalder 1996; Argus 1997). Thus, plant stature and growth form, leaf
morphology, and bud characteristics have been important characters for species identification
(Dorn 1976; Argus 2010; Eckenwalder 2010) despite the high intraspecific variability and
propensity for plasticity of these traits (Wu et al. 2015).
Here we seek to understand the sources of phylogenetic conflict in both Salix and
Populus. Our ASTRAL phylogeny was based on genome-wide sequence capture loci from a
large set of mostly North American and Asian diploid Salix and Populus species. The
phylogeny of a subset of these Salix samples was previously investigated using DNA barcode
markers, but the resulting tree lacked resolution (Percy et al. 2014). A phylogeny of the
majority of the Populus samples was previously analyzed by Wang et al. (2020) using 5305
(Sanderson et al. 2020) successfully reconstructs the species tree. We use our ASTRAL tree
to compare the impact of hybridization on gene tree discordance and the chronology and rates
of diversification across these two genera. The specific goals of our study are to: 1) provide
an integrated phylogenetic hypothesis of the sister genera Populus and Salix, 2) to estimate
the timing and rates of diversifications of major clades within each genus, and 3) to assess the
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
8
association between hybridization history and regions of the phylogeny that exhibit conflict
among gene trees. Finally, we discuss the implications of our results in the context of other
well-known hybridizing groups of species (syngameons) such as oaks (Cannon et al. 2020)
and pines (Buck et al. 2022).
Materials and Methods
Data collection
The 165 samples included in this study were drawn from 71 new collections, 44
herbarium samples, and 50 previously sequenced genomes that represented all five Salix
subgenera, 25 Salix sections, and all five Populus sections (Table S1). All species were
considered diploids based on chromosome counts reported in www.tropicos.org except S.
discolor, which is likely a polyploid, and S. richardsonii, for which there was no information
concerning chromosome counts. The sampled Salix species were primarily native to North
America, but also from Europe and Asia (Table S1). Fourteen Populus samples (7 species)
and 83 Salix samples (45 species) were genotyped using a custom sequence target capture kit
designed for the Salicaceae (Supplemental Methods; Sanderson et al. 2020). An additional 6
outgroup species (Azara dentata, A. integrifolia, A. lanceolata, A. microphylla, Carrierea
calycina, Idesia polycarpa, and Poliothyrsis sinensis), 54 poplar samples (26 species), and 8
Salix samples (7 species) were genotyped using whole genome sequencing. All sequences
were assembled into putatively homologous gene sequences using the HybPiper pipeline
(Johnson et al. 2016). For the whole genome sequences, HybPiper only assembled the loci
included in the target capture array (see Table S2 for coverage statistics). Paralogous
sequences may align to the same reference gene, with the potential to generate trees that are
inconsistent with the species tree or have excessively long branches. To address these
concerns, we removed both copies of 413 sequence capture locus alignments that included
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
9
potential paralogs that were identified using Hybpiper. For each gene sequence, all sites with
>25% gaps (-gt 0.75) and without at least 90% of residues that overlap with at least 90% with
the rest of the sequences (-resoverlap 0.90 -seqoverlap 0.90) were removed using trimal
v1.4.rev22 (Capella-Gutierrez et al. 2009). TreeShrink 1.3.4 was used to detect and filter out
sequences with excessively long branches in each gene tree (Mai et al. 2018). 787 alignments
remained after filtering and removing gene alignments likely containing paralogous
sequences and were used for all downstream phylogenetic analyses (see Supplemental
Methods for additional details; Table S3).
Phylogenetic and dating analyses
Two approaches were used for phylogeny estimation: 1) A two-step approach using
ASTRAL that first estimated trees for each gene and then identified the best species tree
based on minimizing quartet distances among gene trees (Mirarab et al. 2014), and 2)
identifying the maximum likelihood tree based on concatenating all genes in our sample. Full
details of these analyses are provided in the Supplemental Methods. In brief for the two-step
approach, gene trees were estimated using IQTREE 2.0.3 (Nguyen et al. 2015) and both an
ASTRAL tree of all individuals and the ASTRAL species tree (Rabiee et al. 2019) were
inferred using ASTRAL-MP (v 5.12.2 (Yin et al. 2019). In our ASTRAL tree of all
individuals, four species were represented by intraspecific samples that did not cluster
together (Fig. S1): Salix bebbiana, S. eriocephala, S. pseudomonticola, and Populus
ningshanica; these are represented by a postscript 1 & 2 in the figures and Table S1. For this
reason, we used two individuals to represent each of these four species in the ASTRAL
species tree. Both local posterior probability and 100 multilocus bootstraps were calculated
for each ASTRAL tree. For the concatenation approach, a single alignment was constructed
by concatenating all 787 genes (1,058,955 sites). IQTREE was used to estimate the most
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
10
likely concatenated tree using the GTR+F+R10 substitution model. 1000 mulitlocus
bootstraps and SH-alRT tests (Guindon et al. 2010) also were computed for this tree.
A dated species tree was calculated using *BEAST2 (Heled et al. 2010). Because of
the long computation times required for sampling, a gene shopping approach (Smith et al.
2018) was used to select sets of five genes for estimating divergence times (9,264 sites:
SapurV1A.0003s0350, SapurV1A.0045s0240, SapurV1A.0050s0650, SapurV1A.0139s0330,
SapurV1A.0260s0050). These genes were first selected for high consistency with species tree
topologies, minimized root-to-tip variance, and maximal tree length (Smith et al. 2018) and
then screened for consistency of basal nodes with the ASTRAL species tree. Because these
five genes represented <1% of the genes used for generating our species tree, we selected an
additional five genes using the same criteria to assess the consistency of the results across
different gene sets (8,568 sites: SapurV1A.0211s0160, SapurV1A.0789s0070,
SapurV1A.0857s0020, SapurV1A.0900s0040, SapurV1A.1178s0060). See the Supplemental
methods for additional information on gene selection and for *BEAST2 parameter settings.
The calibration date for the root of the tree (divergence between Azara and all others) was
drawn from a normal distribution with a mean of 65.0 Ma and standard deviation of 1.0
(following Wang et al. 2020), and the calibration of the crown clade of Populus + Salix was
drawn from a normal distribution with a mean of 49 Ma and a standard deviation of 1.0
following Percy et al. (2014) and based on the Pseudosalix handleyi fossil (Boucher et al.
2003) from the Eocene Green River formation that has been dated at ca. 49 Ma (Smith et al.
2010). Using a wider standard deviation of 3.0 for the distributions resulted in much larger
variance in the estimates of node ages, but only slight changes in the estimates for divergence
times (Fig. S7).
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
11
Hybridization and gene flow inference
Patterns of current and historical hybridization within the Populus and Salix clades
were estimated using ABBA-BABA, f4, and fbranch analyses (Patterson et al. 2012; Malinsky
et al. 2018) calculated using Dsuite (Malinsky et al. 2021). The fbranch analysis is heuristic
designed to account for phylogenetic correlation among f4-ratio results calculated with
phylogenetically correlated samples. The fbranch metrics assign significance to internal
branches in the phylogeny when excess sharing of alleles that is consistent with hybridization
is found across a clade (Malinsky et al. 2018; Malinsky et al. 2021). We generated separate
VCF files for Salix and Populus using GATK v4.2.6.1 (see Supplemental Methods for
details; McKenna et al. 2010) to calculate f4-ratio and fbranch statistics.
Diversification rate analysis
Two methods were used to address diversification rates across the Salicaceae. First,
Bayesian Analysis of Macroevolutionary Mixtures (BAMM; Rabosky 2014) was used to
identify credible shifts in the diversification rate across lineages with the
expectedNumberOfShifts prior set to 1.0. Because our species-level sampling was not uniform
across genera, we adjusted for non-random incomplete taxon sampling (Table S4). For each
dated tree, two independent MCMC chains using different seeds were run for 10 million
generations each resulting in ESS>200, and the 95% credible set of shift configurations was
calculated after removing 10% for burn-in. Second, branch-specific diversification rates were
estimated using the ClaDS model (Maliet et al. 2019) and calculated using data augmentation
(Maliet et al. 2022).
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
12
Results
Species tree and gene tree conflict
Of the 1216 loci that were targeted, 787 passed filtering criteria with an average of 162
individuals per gene, 1236 sites per sequence, 305.0 parsimony-informative sites, 132.3
singletons, and 908.2 constant sites per gene. Our ASTRAL and concatenated trees strongly
supported monophyly for both Salix and Populus, the monophyly of Salix subgenera Protitea
and Longifoliae, and Populus subgenera Turanga and Populus (Figs. 1, S1, S2; Eckenwalder
1996; Wagner et al. 2018; Wang et al. 2020; Gulyaev et al. 2022). Our phylogenies also
supported recent discoveries of the polyphyly of the Vetrix and Chamaetia subgenera in Salix
(Wagner et al. 2018) and the polyphyly of the Tacamahaca and Aigeiros subgenera in
Populus (Wang et al. 2020). Notably in Salix, our study and that of Wager et al. (2018) each
identified up to 4 independent evolutionary events leading to dwarf willows (subgenus
Chamaetia, Fig. 1), which are prominent components of northern hemisphere arctic and
alpine ecosystems. However, because these studies had little taxonomic overlap, it is difficult
to discern whether we identified the same or different events.
Most of our samples identified as the same species clustered together in both the
ASTRAL tree of individuals and the concatenated tree (Fig. S1). Clades in Populus were
generally more strongly supported by consistency among gene trees than clades in Salix
(Figs. 1, S3); this lack of support was especially apparent in the backbone of the Salix subg.
Chamaetia+Vertix clade where some ASTRAL tree bipartitions were present in 4 or fewer
gene trees. The most basal node supported by 4 or fewer gene trees describes a split between
S. richardsonii and clades 2-7 that was found in only 4 of 787 genes (Fig. S3). The most
common alternative bipartition for this node was present in 13 gene trees and placed both S.
richardsonii and S. lasiandra outside of the bipartition including subgenera
Salix+Longifoliae+Chamaetia/Vetrix (Fig. S4A). In nodes supported by bipartitions in 0 gene
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
13
trees, no bipartition was supported by more than 9 gene trees, indicating weak support for a
consistent alternative topology. The concatenated and ASTRAL trees that included all
individuals also placed Salix lasiandra and S. richardsonii in drastically different positions
(Fig. S1). In Populus, notable alternate bipartitions included one in which 30 gene trees
supported the placement of P. qiongdaoensis outside of subgenus Populus (Fig. S4B) and a
second in which 37 gene trees supported the placement of subgenus Turanga as sister to the
remaining Populus species (Fig. S4C)
Ancient intraspecific gene flow
The extent of biased intraspecific gene flow indicative of hybridization was approximately
the same in Populus and Salix (52% of Dtree and 60% of Dmin were significant after
Benjamini-Hochberg adjustment in Salix vs. 64% of Dtree and 58% of Dmin in Populus; Tables
S5-S8). Because D and f4 statistics across a clade are phylogenetically correlated, we used the
heuristic fbranch analysis based on the ASTRAL species tree to assess the history and timing of
hybridization during the diversification of Populus and Salix (Fig. S5, Table S9, S10). The
number of fbranch statistics that indicated >5% gene biased flow across species due to
hybridization were similar in Salix (6.6% of fbranch values above 5%) and Populus (7.0% of
fbranch values above 5%) but were more commonly associated with deep internal branches of
Salix than Populus (Tables S9, S10). Because many fbranch values were above zero, we focus
here only on those with values >0.05 that were associated with basal nodes. From the patterns
of fbranch values, we inferred hybridization between ancestors of S. lasiandra and the
Chamaetia/Vetrix ancestor (fbranch 0.05; Table S9) and ancestors of subgenus Protitea (arrows
labelled 1 in Figs. 1 & S5; fbranch 0.18-0.25; Table S9). The presence of these ancient
hybridization events is supported by the incongruence among gene trees and the lack of
strong support for the phylogenetic placement of S. lasiandra. Additionally, signals of gene
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
14
flow between S. richardsonii and subgenus Protitea likely contributes to its inconsistent
phylogenetic placement across gene trees (arrow 2 in Figs. 1 & S5; fbranch = 0.088 & 0.113;
Table S9). Although interspecific gene flow is identified in many places during the history of
Salix, we emphasize three additional ancient hybridization events because of their potential to
impact on the lack of consistency among basal nodes among gene trees. First, the fbranch
metrics indicated evidence for gene flow between the ancestors of both S. triandra and S.
arbutifolia and ancestors of the lineage leading to subgenera Longifoliae, Vetrix and
Chamaetia (arrow 3 in Figs. 1 & S5; fbranch = 0.088 & 0.113; Table S9). Second, ancient
geneflow occurred between the ancestors of subgenus Longifoliae with ancestors of clade 2-7
(arrow 4 in Figs. 1 & S5; fbranch = 0.071 to 0.164; Table S9). Third, the fbranch analysis
indicates ancient hybridization and gene flow among ancestors of clades 7 and the ancestors
of subgenus Longifloiae (arrow 5 in Figs. 1 & S5; fbranch = 0.053 to 0.186;; Table S9). In
some cases, however, strong evidence of hybridization may signal the incorrect placement of
the species on the tree. For instance, the lack of evidence of hybridization in S. geyriana and
strong evidence for hybridization between S. sitchensis and an ancestor of clades1-7 suggests
that either S. sitchensis or S. geyriana may be misplaced on the phylogeny and the multiple
signals of hybridization result from a single ancient event. In Populus the fbranch analysis
based on the ASTRAL species tree indicated widespread hybridization among ancestors or
extant members of the two major clades comprising the Populus subgenus (arrows labelled 6
in Figs. 1 & S6; fbranch = 0.120 to 0.259; Table S10) as well as signals of ancient or ongoing
hybridization within the Tacamahaca/Aigeiros clade and the Leucoides grade (arrows
labelled 7 in Figs. 1 & S6; Table S10).
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
15
Divergence times and rates
The two data sets used to estimate the *BEAST2 dated species trees were both large
(>8500 sites) and drawn from genes in the top 7% of our objective criteria. Dating of many
nodes was within the overlap of 95% highest posterior density for the two estimates (HPD;
Fig. 2; Tables S11, S12). For instance, the divergence between Populus and Salix was
P. mexicana from the
remainder of Populus Salix subgenera Vetrix
and Chamaetia
estimates of divergence times differed substantially between the data sets with no overlap in
95% HPD. In Populus these included the timing of divergence of subgenera Turanga (node
Populus , and in Salix these included the
timing of divergence of subgenera Protitea
inconsistency reflects variance among the gene histories within Populus and Salix and
identify regions of the tree where interpreting the ages of divergence require elevated caution.
We also note here that our HPD estimates were constrained by our selection of the standard
deviation of the a priori distributions of cali
uncertainty for node dates increased dramatically (Fig. S7; Tables S11, S13).
Despite slightly different estimates of node ages in our two dated trees, both the
BAMM and ClaDS analyses indicated strong support for similar patterns of shifts in the
diversification rates within both of our dated trees (Fig. 3; Table S14). In both dated trees,
diversification rate increased near the Salix-Populus split and a second increase occurred near
the origination of the Vetrix-Chamaetia clade in Salix (Fig. 3). For the first dated tree, the
BAMM analysis identified two branches with high marginal shift probabilities (Fig. 3A), and
these shifts were largely supported in the ClaDS analysis (Fig. 3C). The BAMM analysis also
identified substantial support for an increase in diversification rate near the Populus-Salix
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
16
split for the second dated tree, with the exception that the model with most support indicated
that Populus subgenus Abaso (with only P. mexicana) retained the same diversification rate
at the outgroups (Fig. 3B). It is notable that the marginal odds ratio supporting the shift in
diversification at the branch that included all of Populus was nearly as high (0.33-0.34) as the
branch that did not include P. mexicana (0.48-0.50), indicating that the shift was likely near
the base of the Salix+Populus clade, but may not have included Populus subgenus Abaso.
The ClaDS analysis also supported a rate shift in this general region of the phylogeny. The
BAMM analysis also identified in the second dated tree substantial support for an increase in
diversification rate in clade that includes the subgenus Longifoliae along with the Vetrix-
Chamaetia clade in Salix (Fig. 3B). Again, this pattern was supported in the ClaDS analysis
(Fig. 3D).
Discussion
Ancient hybridization in willows and poplars
The scale of hybridization on Populus and Salix is remarkable in its taxonomic, genomic, and
chronological extent. Signals of ancient hybridization and introgression across diverged
lineages have persisted in descendant lineages, are particularly evident in Salix, and may
better account for gene tree discordance than contemporary hybridization. Perhaps the most
compelling example of ancient hybridization was the strong signal of biased gene flow
between the ancestors of Salix subgenus Longifoliae and the ancestors of clades 2-7 (depicted
as event 4 on figs. 1 & S5) because of the large numbers of derived species affected by this
event. Our fbranch analysis estimated that approximately 10-20% of the Salix genomes in the
Vetrix-Chamaetia clade were affected by the persistence of genes from ancient hybridization
events, which based on our dated trees, began at least 5 Mya. This pattern is consistent with
the low levels of chloroplast genomic diversity across the Vetrix-Chamaetia subgenera which
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
17
was hypothesized to have been partially influenced by high levels of hybridization and
introgression (Wagner et al. 2021b). Our results indicate that hybridization events affecting
genomic variation in subgenera Salix and Longifoliae likely occurred even earlier, perhaps
nearly 10Mya. These hybridization events have likely contributed to the difficulties in
reconstructing relationships within Salix in the present and previous studies (Leskinen et al.
1999; Barkalov et al. 2014; Percy et al. 2014; Lauron-Moreau et al. 2015; Wu et al. 2015; Liu
et al. 2016). In Populus, signals of hybridization between subgenus Turanga and several
members of Aigieros-Tacamahaca contributed to the low support for the position of subgenus
Turanga clade and may underlie incongruence between the ASTRAL and concatenated trees
(Fig. S1). Nonetheless, caution is required when interpreting the fbranch analysis because it is
dependent on the correct topology of the framework phylogeny. An example of this impact is
evident in the high level of hybridization indicated between S. sitchensis and most other
species in the Vetrix-Chamaetia clade, yet the lack of hybridization between the putative
close relative S. geyeriana and these same taxa. This pattern was unlikely to have resulted
from recent hybridization between S. sitchensis and each one of these species but more likely
resulted from the misplacement of S. sitchensis and/or S. geyeriana in the phylogeny.
Previous studies have reported signals of ancient hybridization in both genera.
Chloroplast capture reflecting hybridization of ancient lineages has been reported in Salix
(Brunsfeld et al. 1992; Hardig et al. 2000; Gulyaev et al. 2022). An ancestral member of
subgenus Longifoliae captured a chloroplast of a member of subgenus Protitea (Gulyaev et
al. 2022) and five cases of arguably more recent chloroplast capture within Salix section
Longifoliae were reported by Brunsfeld et al. (1992). Notably, crossing studies indicate that
extant members of subgenera Longifoliae and Protitea are reproductively isolated (Mosseler
1990), so the hybridization generating the former chloroplast capture event likely occurred
before reproductive isolation evolved. The deep history of hybridization in Salix also is
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
18
reflected in the evidence for gene flow among ancestral lineages in our study and indicates a
long history of concomitant hybridization and speciation in the genus. As others have argued
for Populus (Cronk et al. 2018), this pattern of ongoing speciation with hybridization may be
best represented as a syngameon (Lotsy 1925) and may exhibit emergent properties such as
the ability to draw on elevated levels of standing variation for adaptive evolution (Cannon et
al. 2020; Cannon 2021). Envisioning Salix as a syngameon redefines evolutionary units as
larger combinations of hybridizing species and our data and previous results (Hardig et al.
2000; Murphy et al. 2022) suggest that the Salix syngameon exhibits a complex web of
ongoing hybridization and partial reproductive isolation that has persisted for millions of
years.
Two ancient chloroplast capture events also have been previously identified in
Populus. One of these, which was based on many of the same individuals as in the present
study (Table S1; Wang et al. 2020), discovered that an ancestor of P. heterophylla captured
the chloroplast of P. mexicana ancestors (Liu et al. 2017; Wang et al. 2020). The second
event was the capture of the chloroplast of P. alba ancestors by an ancestor of P. nigra
(Smith et al. 1990). Our data, however, found no evidence of ancient hybridization among
these lineages in the nuclear genomes. This lack of evidence may result from the limited
numbers of genes that we sampled, the limited influence of these ancient hybridization events
on only the chloroplast genomes (Tsitrone et al. 2003), or the lack of power for ABBA-
BABA type tests to detect ancient hybridization. f4 and D-statistics likely underestimate
historical hybridization for several reasons (Jiang et al. 2020; Ji et al. 2022). First, these
statistics only identify significantly biased directional introgression (numbers of ABBA and
BABA differ), but when bi-directional introgression is similar, power to discriminate from
incomplete lineage sorting is low. Second, ancient deviation between ABBA and BABA site
patterns will be eroded as sites accumulate multiple mutations on the tree. Finally, the
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
19
ABBA-BABA model cannot detect hybridization between sister lineages, including ancient
sister lineages.
Although our recovered topology for Populus was similar to Wang et al. (2020), it
was not identical. A notable difference between our tree and that of Wang et al. (2020) is that
our tree placed P. angustifolia as sister to the P. trichocarpa-P. balsamifera clade, whereas
the Wang et al. (2020) placed P. angustifolia at the base of a larger clade including multiple
North American and Asian species. Importantly, the placement of Wang et al. (2020)
indicates that either P. szechuanica, P. laurifolia, and P. koreana or P. angustifolia may have
speciated due to vicariance or long-distance dispersal from North America to Asia and the
divergence of P. angustifolia predates this event. The difference in the placement of P.
angustifolia is particularly interesting and worthy of further study because it commonly
hybridizes with both P. trichocarpa and P. balsamifera (Brayshaw 1965; Chhatre et al.
2018), which may influence patterns of gene tree coalescence, and the perceived relationships
among species. Interestingly, some foliar pathogens of P. angustifolia are related to foliar
pathogens of Asian members of Tacamahaca and absent in North American members (Busby
et al. 2012), lending support to the hypothesis of a recent trans-Beringian migration of P.
angustifolia.
Divergence times and rates
The estimates for divergence times among subgenera within Salix and Populus
presented here must be considered tentative because we did not calibrate internal nodes in
each genus. Many fossils of Salix and Populus from both Asia and North America have been
identified (Collinson, M.E. 1992), but it has been difficult to accurately assign them to extant
taxonomic groups without robust phylogenies. We chose to rely primarily on the molecular
clock to estimate diversification dates within these genera instead of including internal
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
20
calibrations following Percy et al. (2014) and Zhang et al. (2018). We note that the estimates
of diversification dates by Wu et al. (2015) were much earlier than we report here and
included an internal Salix late-Oligocene (ca. 26Ma) leaf fossil to calibrate of the origination
of subgenus Vetrix (Wolfe pers. comm. 1991 in Collinson 1992). Also, Populus fossils
hypothesized to belong to subgenus Tacamahaca are described from the Late Oligocene
Creede flora of Colorado (Wolfe et al. 1990; Collinson 1992), a date much earlier than our
estimate of the diversification of this subgenus. However, our and other recent phylogenies
have identified both subgenus Tacamacaha and Vetrix as polyphyletic (Wagner et al. 2018;
Wang et al. 2020), suggesting that the characters used to categorize the fossils should be
reexamined. The last thorough review of Salicaceous fossils was published in 1992
(Collinson 1992), and updating of this group in relation to the most up-to-date phylogenies
would aid greatly in developing a more accurate estimate of diversification times in the
Salicaceae.
We estimated divergence times based on two different sets of 5 genes selected using a
gene shopping approach (Smith et al. 2018). Inconsistency among gene trees and signals of
past interspecific gene flow, as in our data set (Fig. S3) indicates that ILS or post-divergence
hybridization is common in Salix and Populus. Interspecific gene flow is known to result in
underestimating divergence times (Leaché et al. 2013; Tiley et al. 2023). Our approach
minimized the influence of hybridization by selecting sets of genes with the most consistency
with the gene tree. Selection of a significantly larger set of gene trees will result in the
inclusion of genes with larger influences of hybridization and thus result in increasing
underestimates of divergence times. Simultaneously estimating historical hybridization and
divergence times using a multispecies-coalescent-with-introgression model (Tiley et al. 2023)
is one approach to solving this problem, but given the large computational demands this
approach is better suited for a smaller set of taxa than addressed here.
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
21
In Salix we identified a burst of diversification near the origin of the Vetrix-
Chamaetia clade that likely increased the level of incomplete lineage sorting (ILS) and
contributed to the lack of support for inferred relationships within this clade (Roch et al.
2015). We also identified a second increase in the diversification rate near the divergence of
Populus from Salix. The mechanisms driving these increases in diversification rate remain
speculative. Although these shifts are accompanied with changes in seed dispersal,
pollination vectors, and growth form (Argus 2010; Eckenwalder 2010), and even a shift from
an XY to a ZW sex determination system in Salix (Gulyaev et al. 2022; Hu et al. 2022), each
of these shifts occurs only once or twice on the tree, so too few phylogenetically independent
events have occurred for powerful statistical tests of association. We suspect that detailed
studies of the influence of insect pollination on reproductive isolation and genetic structure in
Salix may help to address these mechanisms as insect pollination influences diversification
rates in other systems (Wessinger 2021). Finally, these patterns may be useful for meta-
analyses including a larger set of taxonomic groups, or future studies may find more detailed
patterns associated with genetic or morphological changes that can shed light on the drivers
of diversification in these two important genera.
Future directions and conclusions
Our sample of Populus (31 of ~100 species) included the majority of diversity that
occurs across Asia, Europe, Africa, and North America, but our sampling of Salix (46 of
~500 species) was limited to primarily on diploids (one tetraploid) and focused mainly on
North America but included a few Eurasian species (Table S1). The Eurasian samples in
Salix were interspersed with the North American species in clades 3 & 7 within subgenus
Vetrix/Chamaetia as well as in subgenus Salix and Protitea. Given this pattern, it is likely
that the increased speciation rate in Vetrix/Chamaetia is not geographically biased. A direct
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
22
test of this hypothesis would be timely. Addition of more Eurasian species as well as
polyploids will be revealing both in relation to relationships among taxonomic groups as well
as sources of trans-specific gene flow.
The continued development of analytical methods for quantifying ancient
interspecific gene flow will permit investigations into the prevalence of ancient hybridization,
its impacts on adaptation, and its biological and phylogenetic correlates. Ancient
hybridization inferred from chloroplast capture has been identified in large numbers of plants
groups (Rieseberg et al. 1991). However, chloroplast capture may result from unique
properties of the hybridizing species (Tsitrone et al. 2003) and may not be present in all
taxonomic groups affected by ancient hybridization. Interspecific gene flow is usually not
categorized as either contemporary or ancient, but the relative influences of ancient versus
recent introgression events may provide insight into the sources of different classes of genetic
variants under selection (e.g. Menon et al. 2021). Recent studies in Quercus, Pinus, and
Populus have documented adaptive introgression among hybridizing groups of multiple
species (Chhatre et al. 2018; Leroy et al. 2020; Buck et al. 2022). Along with Populus and
Salix, ancient gene flow has also been identified in Quercus, where phylogenetic conflict in
chloroplast genomes suggests hybridization during early diversification (Yang et al. 2021)
indicating at least two plant families with a combination of both contemporary and ancient
interspecific gene flow. Understanding patterns of ancient interspecific gene flow in relation
to contemporary hybridization may provide further insight into biological factors correlated
with hybridization (Mitchell et al. 2019) and factors associated with the development of
syngameons.
Examples of the extended effects of contemporary hybridization in Salix and Populus
are well-known (Hardig et al. 2000; Evans et al. 2008; Lexer et al. 2010). The cumulative
effects of persistent hybridization over eons as diversification unfolded, however, may result
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
23
in different qualities of adaptation and diversification than isolated cases of contemporary
hybridization. In Salix and Populus this history has resulted in a tangle of gene histories
within each genus, with some clades having developed monophyly and others that may never
be resolved. Hybridization also may have contributed changes in the diversification rate of
species in the Vetrix-Chamaetia clade of Salix. Understanding the characteristics associated
with and generated from long-term and persistent interspecific gene flow will elucidate
whether these properties confer fundamentally different pattens of adaptation and speciation
(Cannon 2021).
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
24
Supplementary Methods, Tables, and Figures
Supplementary Methods, Tables, and Figures are available at
https://datadryad.org/stash/dataset/doi:10.5061/dryad.41ns1rnf7?
Data Accessibility
All raw data are available at NCBI and accessions are listed in Table S1. All gene alignments
are available on Dryad at https://datadryad.org/stash/dataset/doi:10.5061/dryad.kprr4xh84
Acknowledgements
We thank Peter Zhelev and Gancho Slavov for S. triandra collections, Andrew Hamstetter for
insightful discussions of diversification analyses, and Pascal Title for computational assistance with
BAMM. This research was supported by grants from the US National Science Foundation (1542509,
1542599, 1542479, 1542486) and the National Natural Science Foundation of China (31561123001).
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
25
Literature Cited
Argus GW. 1997. Infrageneric classification of Salix (Salicaceae) in the new world.
Systematics Botany Monographs, 52:1-121.
Argus GW. 2010. Salix L. In: Committee FoNAE editor. Flora of North America North of
Mexico [Online]. New York and Oxford.
Barkalov VY, Kozyrenko MM. 2014. Phylogenetic relationships of Salix L. subg. Salix
species (Salicaceae) according to sequencing data of intergenic spacers of the chloroplast
genome and ITS rDNA. Russian Journal of Genetics, 50:828-837.
Boeckler GA, Gershenzon J, Unsicker SB. 2011. Phenolic glycosides of the Salicaceae and
their role as anti-herbivore defenses. Phytochemistry, 72:1497-1509.
Boucher LD, Manchester SR, Judd WS. 2003. An extinct genus of Salicaceae based on twigs
with attached flowers fruits, and foliage from the Eocene Green River Formation of Utah and
Colorado, USA. Am J Bot, 90:1389-1399.
Brayshaw TC. 1965. Native poplars of southern Alberta and their hybrids. Ottowa, CA,
Brunsfeld SJ, Soltis DE, Soltis PS. 1992. Evolutionary patterns and processes in Salix sect
Longifoliae - evidence from chloroplast DNA. Systematic Botany, 17:239-256.
Buck R, Ortega-Del Vecchyo D, Gehring C, Michelson R, Flores-Renteria D, Klein B,
Whipple AV, Flores-Renteria L. 2022. Sequential hybridization may have facilitated
ecological transitions in the Southwestern pinyon pine syngameon. New Phytol.
Busby PE, Aime MC, Newcombe G. 2012. Foliar pathogens of Populus angustifolia are
consistent with a hypothesis of Beringian migration into North America. Fungal Biology,
116:792-801.
Cannon CH. 2021. Is speciation an unrelenting march to reproductive isolation? Molecular
Ecology, 30:4349-4352.
Cannon CH, Petit RJ. 2020. The oak syngameon: more than the sum of its parts. New Phytol,
226:978-983.
Capella-Gutierrez S, Silla-Martinez JM, Gabaldon T. 2009. trimAl: a tool for automated
alignment trimming in large-scale phylogenetic analyses. Bioinformatics, 25:1972-1973.
Chhatre VE, Evans LM, DiFazio SP, Keller SR. 2018. Adaptive introgression and
maintenance of a trispecies hybrid complex in range-edge populations of Populus. Molecular
Ecology, 27:4820-4838.
Collinson ME. 1992. The early fossil history of Salicaceae - A brief review. Proceedings of
the Royal Society of Edinburgh Section B-Biological Sciences, 98:155-167.
Cronk QC, Needham I, Rudall PJ. 2015. Evolution of catkins: Inflorescence morphology of
selected Salicaceae in an evolutionary and developmental context. Front Plant Sci, 6:1030.
Cronk QC, Suarez-Gonzalez A. 2018. The role of interspecific hybridization in adaptive
potential at range margins. Molecular Ecology, 27:4653-4656.
Degnan JH, Rosenberg NA. 2009. Gene tree discordance, phylogenetic inference and the
multispecies coalescent. Trends Ecol Evol, 24:332-340.
Desborough MJR, Keeling DM. 2017. The aspirin story - from willow to wonder drug.
British Journal of Haematology, 177:674-683.
Dorn RD. 1976. A synopsis of American Salix. Canadian Journal of Botany, 54:2769-2789.
Durand EY, Patterson N, Reich D, Slatkin M. 2011. Testing for ancient admixture between
closely Related populations. Mol Biol Evol, 28:2239-2252.
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
26
Eckenwalder JE. 1996. Systematics and Evolution of Populus In: Stettler RF, Bradshaw HD,
Heilman PE, T.M.Hinckley editors. Biology of Populus and its Implications for Management
and Conservation. Ottawa, Canada, NRC Research Press, p. 7-56.
Eckenwalder JE. 2010. Populus L. In: Committee FoNAE editor. Flora of North America
North of Mexico [Online], New York and Oxford.
Edwards SV. 2009. Is a new and general theory of molecular systematics emerging?
Evolution, 63:1-19.
Evans LM, Allan GJ, Shuster SM, Woolbright SA, Whitham TG. 2008. Tree hybridization
and genotypic variation drive cryptic speciation of a specialist mite herbivore. Evolution,
62:3027-3040.
Friedman J, Barrett SCH. 2009. Wind of change: new insights on the ecology and evolution
of pollination and mating in wind-pollinated plants. Annals of Botany, 103:1515-1527.
Grant V. 1981. Plant Speciation. New York, Columbia University Press.
Green RE, Krause J, Briggs AW, Maricic T, Stenzel U, Kircher M, Patterson N, Li H, Zhai
WW, Fritz MHY, et al. 2010. A draft sequence of the Neandertal genome. Science, 328:710-
722.
Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W, Gascuel O. 2010. New
algorithms and methods to estimate maximum-likelihood phylogenies: Assessing the
performance of PhyML 3.0. Systematic Biology, 59:307-321.
Gulyaev S, Cai XJ, Guo FY, Kikuchi S, Applequist WL, Zhang ZX, Horandl E, He L. 2022.
The phylogeny of Salix revealed by whole genome re-sequencing suggests different sex-
determination systems in major groups of the genus. Annals of Botany, 129:485-498.
Hardig TM, Brunsfeld SJ, Fritz RS, Morgan M, Orians CM. 2000. Morphological and
molecular evidence for hybridization and introgression in a willow (Salix) hybrid zone.
Molecular Ecology, 9:9-24.
Heled J, Drummond AJ. 2010. Bayesian inference of species trees from multilocus data. Mol
Biol Evol, 27:570-580.
Hu N, Sanderson BJ, Guo M, Feng G, Gambhir D, Hale H, Wang D, Hyden B, Liu J, Ma T,
et al. 2022. An unusual origin of a ZW sex chromosome system. PREPRINT (Version 1)
available at Research Square.
Hudson RR. 1983. Testing the Constant-Rate Neutral Allele Model with Protein-Sequence
Data. Evolution, 37:203-217.
Ji J, Jackson DJ, Leaché AD, Yang Z. 2022. Power of Bayesian and heuristic tests to detect
cross-species introgression with reference to gene flow in the Tamias quadrivittatus group of
North American chipmunks. Systematic Biology.
Jiang XD, Edwards SV, Liu L. 2020. The Multispecies Coalescent Model Outperforms
Concatenation Across Diverse Phylogenomic Data Sets. Systematic Biology, 69:795-812.
Johnson MG, Gardner EM, Liu Y, Medina R, Goffinet B, Shaw AJ, Zerega NJC, Wickett NJ.
2016. HybPiper: Extracting coding sequence and introns for phylogenetics from high-
throughput sequencing reads using target enrichment. Applications in Plant Sciences, 4.
Karrenberg S, Kollmann J, Edwards PJ. 2002. Pollen vectors and inflorescence morphology
in four species of Salix. Plant Systematics and Evolution, 235:181-188.
Keefover-Ring K, Carlson CH, Hyden B, Azeem M, Smart LB. 2022. Genetic mapping of
sexually dimorphic volatile and non-volatile floral secondary chemistry of a dioecious
willow. Journal of Experimental Botany, 73:6352-6366.
Lauron-Moreau A, Pitre FE, Argus GW, Labrecque M, Brouillet L. 2015. Phylogenetic
Relationships of American Willows (Salix L., Salicaceae). Plos One, 10.
Leaché AD, Harris RB, Rannala B, Yang Z. 2013. The Influence of Gene Flow on Species
Tree Estimation: A Simulation Study. Systematic Biology, 63:17-30.
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
27
Leroy T, Louvet J-M, Lalanne C, Le Provost G, Labadie K, Aury J-M, Delzon S, Plomion C,
Kremer A. 2020. Adaptive introgression as a driver of local adaptation to climate in European
white oaks. New Phytol, 226:1171-1182.
Leskinen E, Alstrom-Rapaport C. 1999. Molecular phylogeny of Salicaceae and closely
related Flacourtiaceae: evidence from 5.8 S, ITS 1 and ITS 2 of the rDNA. Plant Systematics
And Evolution, 215:209-227.
a D, Castiglione S, Fay MF,
Buerkle CA. 2010. Genomic admixture analysis in European Populus spp. reveals
unexpected patterns of reproductive isolation and mating. Genetics, 186:699-712.
Liu X, Wang ZS, Shao WH, Ye ZY, Zhang JG. 2017. Phylogenetic and Taxonomic Status
Analyses of the Abaso Section from Multiple Nuclear Genes and Plastid Fragments Reveal
New Insights into the North America Origin of Populus (Salicaceae). Frontiers in Plant
Science, 7.
Liu X, Wang ZS, Wang DS, Zhang JG. 2016. Phylogeny of Populus-Salix (Salicaceae) and
their relative genera using molecular datasets. Biochemical Systematics and Ecology, 68:210-
215.
Lotsy JP. 1925. Species of linneon. Genetica, 7:487-506.
Mai U, Mirarab S. 2018. TreeShrink: fast and accurate detection of outlier long branches in
collections of phylogenetic trees. Bmc Genomics, 19.
Maliet O, Hartig F, Morlon H. 2019. A model with many small shifts for estimating species-
specific diversification rates. Nature Ecology & Evolution, 3:1086-1092.
Maliet O, Morlon H. 2022. Fast and accurate estimation of species-specific diversification
rates using data augmentation. Systematic Biology, 71:353-366.
Malinsky M, Matschiner M, Svardal H. 2021. Dsuite - Fast D-statistics and related admixture
evidence from VCF files. Molecular Ecology Resources, 21:584-595.
Malinsky M, Svardal H, Tyers AM, Miska EA, Genner MJ, Turner GF, Durbin R. 2018.
Whole-genome sequences of Malawi cichlids reveal multiple radiations interconnected by
gene flow. Nature Ecology & Evolution, 2:1940-1955.
McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, Garimella K,
Altshuler D, Gabriel S, Daly M, et al. 2010. The Genome Analysis Toolkit: A MapReduce
framework for analyzing next-generation DNA sequencing data. Genome Research, 20:1297-
1303.
McVay JD, Hipp AL, Manos PS. 2017. A genetic legacy of introgression confounds
phylogeny and biogeography in oaks. P Roy Soc B-Biol Sci, 284.
Meeuse ADJ. 1975. Taxonomic relationships of Salicaceae and Flacourtiaceae: their bearing
on interpretative floral morphology and dilleniid phylogeny. Acta Botanica Neerlandica,
24:437-457.
Menon M, Bagley JC, Page GFM, Whipple AV, Schoettle AW, Still CJ, Wehenkel C,
Waring KM, Flores-Renteria L, Cushman SA, et al. 2021. Adaptive evolution in a conifer
hybrid zone is driven by a mosaic of recently introgressed and background genetic variants.
Communications Biology, 4:160.
Mirarab S, Reaz R, Bayzid MS, Zimmermann T, Swenson MS, Warnow T. 2014. ASTRAL:
genome-scale coalescent-based species tree estimation. Bioinformatics, 30:I541-I548.
Mitchell N, Campbell LG, Ahern JR, Paine KC, Giroldo AB, Whitney KD. 2019. Correlates
of hybridization in plants. Evol Lett, 3:570-585.
Morales-Briones DF, Kadereit G, Tefarikis DT, Moore MJ, Smith SA, Brockington SF,
Timoneda A, Yim WC, Cushman JC, Yang Y. 2020. Disentangling sources of gene tree
discordance in phylogenomic data sets: Testing ancient hybridizations in Amaranthaceae s.l.
Systematic Biology, 70:219-235.
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
28
Mosseler A. 1990. Hybrid performance and species crossability relationships in willows
(Salix). Canadian Journal of Botany-Revue Canadienne De Botanique, 68:2329-2338.
Murphy EK, Cappa EP, Soolanayakanahally RY, El-Kassaby YA, Parkin IAP, Schroeder
WR, Mansfield SD. 2022. Unweaving the population structure and genetic diversity of
Canadian shrub willow. Sci Rep-Uk, 12.
Nguyen LT, Schmidt HA, von Haeseler A, Minh BQ. 2015. IQ-TREE: A fast and effective
stochastic algorithm for estimating maximum-likelihood phylogenies. Mol Biol Evol, 32:268-
274.
Parins-Fukuchi C, Stull GW, Smith SA. 2021. Phylogenomic conflict coincides with rapid
morphological innovation. Proceedings of the National Academy of Sciences of the United
States of America, 118.
Patterson N, Moorjani P, Luo YT, Mallick S, Rohland N, Zhan YP, Genschoreck T, Webster
T, Reich D. 2012. Ancient admixture in human history. Genetics, 192:1065-+.
Pease JB, Hahn MW. 2013. More accurate phylogenies inferred from low-recombination
regions in the presence of incomplete lineage sorting. Evolution, 67:2376-2384.
Percy DM, Argus GW, Cronk QC, Fazekas AJ, Kesanakurti PR, Burgess KS, Husband BC,
Newmaster SG, Barrett SC, Graham SW. 2014. Understanding the spectacular failure of
DNA barcoding in willows (Salix): does this result from a trans-specific selective sweep?
Mol Ecol, 23:4737-4756.
Philippe RN, Bohlmann J. 2007. Poplar defense against insect herbivores. Canadian Journal
of Botany-Revue Canadienne De Botanique, 85:1111-1126.
Rabiee M, Sayyari E, Mirarab S. 2019. Multi-allele species reconstruction using ASTRAL.
Molecular Phylogenetics and Evolution, 130:286-296.
Rabosky DL. 2014. Automatic detection of key innovations, rate shifts, and diversity-
dependence on phylogenetic trees. Plos One, 9.
Rieseberg LH, Soltis DE. 1991. Phylogenetic consequences of cytoplasmic gene flow in
plants. Evolutionary Trends in Plants, 5:65-84.
Roch S, Steel M. 2015. Likelihood-based tree reconstruction on a concatenation of aligned
sequence data sets can be statistically inconsistent. Theoretical Population Biology, 100:56-
62.
Rohwer J, Kubitzki K. 1984. Salix martiana, a regularly hermaphrodite willow. Plant
Systematics and Evolution, 144:99-101.
Sacchi CF, Price PW. 1988. Pollination of the arroyo willow, Salix-lasiolepis - role of insects
and wind. Am J Bot, 75:1387-1393.
Sanderson BJ, DiFazio SP, Cronk QCB, Ma T, Olson MS. 2020. A targeted sequence capture
array for phylogenetics and population genomics in the Salicaceae. Applications in Plant
Sciences, 8.
Schweitzer JA, Bailey JK, Rehill BJ, Martinsen GD, Hart SC, Lindroth RL, Keim P,
Whitham TG. 2004. Genetically based trait in a dominant tree affects ecosystem processes.
Ecology Letters, 7:127-134.
Shu Y. 1999. Populus Linnaeus. . In: Song H editor. Flora of China @ eFloras. Missouri
Botanical Garden, St. Louis, MO, USA and Harvard UNiversity Herbaria, Cambridge, MA,
USA, p. 139-162.
Smith ME, Chamberlain KR, Singer BS, Carroll AR. 2010. Eocene clocks agree: Coeval Ar-
40/Ar-39, U-Pb, and astronomical ages from the Green River Formation. Geology, 38:527-
530.
Smith RL, Sytsma KJ. 1990. Evolution of Populus-nigra (sect Aigeiros) - introgressive
hybridization and the chloroplast contribution of Populus alba (sect Populus). Am J Bot,
77:1176-1187.
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
29
Smith SA, Brown JW, Walker JF. 2018. So many genes, so little time: A practical approach
to divergence-time estimation in the genomic era. Plos One, 13.
Stettler RF, Bradshaw HD, Heilman PE, Hinckley TM. 1996. Biology of Populus and its
Implications for Management and Conservation. Ottawa, Canada, NRC Research Press.
Tamura S, Kudo G. 2000. Wind pollination and insect pollination of two temperate willow
species, Salix miyabeana and Salix sachalinensis. Plant Ecology, 147:185-192.
Tiley GP, Flouri T, Jiao X, Poelstra JW, Xu B, Zhu T, Rannala B, Yoder AD, Yang Z. 2023.
Estimation of species divergence times in presence of cross-species gene flow. Systematic
Biology.
Tsai CJ, Harding SA, Tschaplinski TJ, Lindroth RL, Yuan YN. 2006. Genome-wide analysis
of the structural genes regulating defense phenylpropanoid metabolism in Populus. New
Phytol, 172:47-62.
Tsitrone A, Kirkpatrick M, Levin DA. 2003. A model for chloroplast capture. Evolution,
57:1776-1782.
Wagner ND, Gramlich S, Horandl E. 2018. RAD sequencing resolved phylogenetic
relationships in European shrub willows (Salix L. subg. Chamaetia and subg. Vetrix) and
revealed multiple evolution of dwarf shrubs. Ecology and Evolution, 8:8243-8255.
Wagner ND, He L, Horandl E. 2020. Phylogenomic relationships and evolution of polyploid
Salix species revealed by RAD sequencing data. Frontiers in Plant Science, 11.
Wagner ND, He L, Horandl E. 2021a. The evolutionary history, diversity, and ecology of
willows (Salix L.) in the European Alps. Diversity-Basel, 13.
Wagner ND, Volf M, Hoerandl E. 2021b. Highly diverse shrub willows (Salix L.) share
highly similar plastomes. Frontiers in Plant Science, 12.
Wang MC, Zhang L, Zhang ZY, Li MM, Wang DY, Zhang X, Xi ZX, Keefover-Ring K,
Smart LB, DiFazio SP, et al. 2020. Phylogenomics of the genus Populus reveals extensive
interspecific gene flow and balancing selection. New Phytol, 225:1370-1382.
Wang YC, Huang J, Li EZ, Xu SJ, Zhan ZF, Zhang XJ, Yang ZQ, Guo FY, Liu KJ, Liu D, et
al. 2022. Phylogenomics and biogeography of Populus based on comprehensive sampling
reveal deep-Level relationships and multiple intercontinental dispersals. Frontiers in Plant
Science, 13.
Wessinger CA. 2021. From pollen dispersal to plant diversification: genetic consequences of
pollination mode. New Phytol, 229.
Wolfe JA, Schorn HE. 1990. Taxonomic revision of the Spermatopsida of the Oligocene
Creede Flora, Southern Colorado., United States Geological Bulletin, p. 1-40.
Wu CI. 1991. Inferences of species phylogeny in relation to segregation of ancient
polymorphisms. Genetics, 127:429-435.
Wu J, Nyman T, Wang D-C, Argus GW, Yang Y-P, Chen J-H. 2015. Phylogeny of Salix
subgenus Salix s.l. (Salicaceae): delimitation, biogeography, and reticulate evolution. BMC
evolutionary biology, 15:311-311.
Yang YY, Qu XJ, Zhang R, Stull GW, Yi TS. 2021. Plastid phylogenomic analyses of
Fagales reveal signatures of conflict and ancient chloroplast capture. Molecular
Phylogenetics and Evolution, 163.
Yin J, Zhang C, Mirarab S. 2019. ASTRAL-MP: scaling ASTRAL to very large datasets
using randomization and parallelization. Bioinformatics, 35:3961-3969.
Young AD, Gillung JP. 2020. Phylogenomics principles, opportunities and pitfalls of big-
data phylogenetics. Systematic Entomology, 45:225-247.
Zhang L, Xi ZX, Wang MC, Guo XY, Ma T. 2018. Plastome phylogeny and lineage
diversification of Salicaceae with focus on poplars and willows. Ecology and Evolution,
8:7817-7823.
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
30
Figures
Figure 1. ASTRAL species tree constructed using 787 genetrees. Numbered arrows represent inferred
hybridization events described in the text. Values above branches indicate posterior probability
support and values below branches indicate bootstrap support. Shades indicate taxon associations
with historically identified subgenera.
Figure 2. Comparison of dated trees estimated from two different sets of 5 genes using by *BEAST2.
A) first set of 5 genes, which was the best set according to criteria described in the methods. B)
second set of 5 genes, which was the second-best set. Stars indicate calibration nodes. Node bars
represent 95% highest posterior density in node height. Subgenera are colored as in Fig. 1. Letters at
nodes are discussed in the text.
Figure 3. Diversification rate changes across the Salicaceae. Panels A and B show the best estimates
of the rate shift set estimated using BAMM for the first and second dated trees, respectively. 60%
and 61% (replicate 1 & 2) of samples in the posterior were assigned the the shift configuration
shown in A, whereas 40% and 42% (replicate 1 & 2) of samples in the posterior were assigned the
the shift configuration shown in B. The 95% credible shift sets for both trees and replicates are
presented in Figures S8 and S9. Shades in A and B represent averaged lineage diversification rates
according to the scales associated with each panel. Numbers at nodes represent the marginal odds
ratios for a shift along branches associated with rate shifts; estimates from the two replicates are
presented separated by a comma. Panels C and D show the ClaDS branch-specific rate estimates for
the first and second dated trees, respectively. The scale represents estimate branch specific
speciation rates.
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
31
Figure 1
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
32
Figure 2
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
Accepted Manuscript
33
Figure 3
Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syad042/7224224 by guest on 16 July 2023
