香京julia种子在线播放

Of the 462 primate species recognized by the NCBI taxonomy resource (Schoch et al., 2020), two monkeys (Cheracebus medemi and Callicebus oenanthe) and five lemurs (Cheirogaleus andysabini, Cheirogaleus grovesi, Cheirogaleus minusculus, Cheirogaleus shethi, and Hapalemur gilberti) have no molecular data annotated as a gene in NCBI GenBank. Thus, 455 primates remain, for which either published molecular phylogenies or sequences accessed to GenBank are available. These would be the target species for our supertree of all primates.

We began synthesizing these 455 primates with the nuclear genomic phylogeny found in Kuderna et al. (2024), which included 205 of our target primate species. We then acquired a primate phylogeny from TimeTree, a phylogenetic database synthesizing 4,185 published molecular phylogenies, including 148,876 species (Kumar et al., 2022). We found 200 of the remaining primate species among these, leaving a further 50. Of these, 26 missing species were found in a recently published timetree of apes and monkeys (Craig et al., 2023b).

For the remaining 24 missing primate species, we conducted a multifaceted literature search for publications containing phylogenetic trees. First, we identified the source studies for genetic samples from these species deposited in GenBank (Clark et al., 2016), assuming these may have been used to build molecular phylogenies. We also searched Google Scholar for any mention of these species in a phylogenetic context, using the same approach employed by TimeTree (Kumar et al., 2022), but with the benefit of a target species list and without limiting our search to timed phylogenies. This yielded seven published primate phylogenies containing 20 species (Springer et al., 2012; Lei et al., 2017; Masters et al., 2017; Salmona et al., 2018; Sgarlata et al., 2019; Hagemann et al., 2022; Blair et al., 2023) (Supplementary Table S1).

However, acquiring these phylogenies in a Newick format for synthesis was not trivial. While many journals endeavor for extensive data availability, this often does not extend to the final results and phylogenies printed in a research article. Of the seven new studies we identified in our search, only one (Lei et al., 2017) had their final timetree available as a standard Newick tree file in the supplementary information. We manually created Newick trees for the remaining six based on phylogeny figures. MEGA’s manual tree drawing tool (Tamura et al., 2021) was used to draw species relationships. Each branch’s length was set to the one measured using ImageJ (Schneider et al., 2012) from the published phylogeny. Finally, we visually inspected each tree and corrected any discrepancies by manually editing the resulting Newick string to ensure accurate reproduction of the published tree figure. These primary timetree files are available in the Supplementary Material.

Among these seven new trees, three (Springer et al., 2012; Lei et al., 2017; Masters et al., 2017) had been time-calibrated by their original authors, so they were used directly. For four others, we had phylogenies where the length of each branch represented the genetic distance (number of substitutions per site). This precluded adding them to our super-timetree directly, so we obtained the literature–consensus secondary calibration time for a given node in each tree from the TimeTree database following (Craig et al., 2023a; Craig et al., 2023b). Then, we constructed an unbiased uniform probability distribution between the upper and lower confidence intervals provided by TimeTree. This phylogeny was then scaled to time using the RelTime (Tamura et al., 2012; Tamura et al., 2018) approach in MEGA (Tamura et al., 2021). All the tree files and calibration schemes are available in the Supplementary Material.

Inclusion of these timetrees accounted for 451 primates of the 455 present in our target set, leaving just four for which no phylogenetic trees were found in the literature: Lepilemur mitsinjoensis, Nycticebus hilleri, Phaner furcifer, and Xanthonycticebus pygmaeus. NCBI GenBank contained sequence data from mitochondrial proteins for each of these species: NADH subunit 3 for Lepilemur mitsinjoensis, NADH subunit 4 for N. hilleri, CYTB for P. furcifer, and NADH subunits 4 and 5 for X. pygmaeus. Using NCBI smartBLAST, we identified GenBank accessions for these four proteins in ten or more additional closely related lemurs and exported alignments for each from GenPept. For each of the four resulting alignments, we trimmed any extra loci preceding or following the protein of interest so that all sequences covered the same range of amino acid positions. Then, we visually inspected the alignments in MEGA (Tamura et al., 2021), built a phylogeny from each alignment in MEGA (Tamura et al., 2021) using the JTT substitution model with little bootstraps (Sharma and Kumar, 2021) as a test of confidence at each node, and timed these trees using RelTime (Tamura et al., 2012; Tamura et al., 2018). All alignments, calibration schemes, and intermediary tree files are available in the Supplementary Material.

Finally, we used Chrono-STA (Barba-Montoya et al., 2024) to combine all the timetrees, including the PG and TT trees, published timetrees, newly timed phylogenies, and timetrees assembled from new alignments. Chrono-STA (Barba-Montoya et al., 2024) combines timetrees based on divergence times between species. This yielded the most taxonomically complete time-calibrated primate phylogeny to date in which every tip and node age is informed by molecular data (Figure 2). It incorporates 455 primates, 98.4% of all those present in the NCBI taxonomic resource, and 100% of those with appropriate molecular data.

We recover the root of the phylogeny, covering the divergence between Haplorhini (apes and monkeys, 316 species) and Strepsirrhini (lemurs and lorises, 139 species) at 71.3 million years ago (mya). This is consistent with the time reported by TimeTree as the consensus of 48 research articles published since 1991 (71.4–77.5 mya). We estimate the crown age of Strepsirrhini as 57.0 mya, the crown age of Haplorhini as 66.2 mya, and the crown age of Simiiformes (apes and monkeys, or Haplorhini minus the tarsiers, 307 species) as 42.2 mya.

This comprehensive molecular timetree of Primates is used to test macroevolutionary hypotheses. We use this phylogeny to compare two alternate explanations for the origin of hyper-diverse clades. First, more speciose clades may simply be older than their less diverse counterparts, allowing greater time for species to accumulate (Coyne and Orr, 1998; McPeek and Brown, 2007; Hedges et al., 2015; Henao Diaz et al., 2019). Second, they may have a faster rate of speciation due to their intrinsic characteristics, such as anatomical features tailored to their habitat, diet, or life history (Sanderson and Donoghue, 1994; Wilson, 2003; Fontanillas et al., 2007; Boucher et al., 2017; Sayol et al., 2019; Han et al., 2020). We compare these two hypotheses across the whole primate phylogeny plus four major lineages of primates: the Strepsirrhini and Haplorhini, plus the two haplorhine clades, Platyrrhini and Catarrhini. While some phylogenetic nesting is inevitable in these results, the phylogenies of each of these five clades should nonetheless be comparable for our purposes.

To test these two alternate hypotheses, we used a pair of Bayesian macroevolutionary models (Supplementary Table S2). First, the cladogenetic diversification rate shift (ClaDS) model infers rates of speciation for each lineage individually, assuming inheritance of the maternal rate with some stochasticity (σ) at each divergence event, which produces an overall trend in speciation (α) for a given clade (Maliet et al., 2019; Maliet and Morlon, 2021). For a clade evolving at a constant rate through time, we expect to see an α near 1.0 and a low σ. By contrast, the TESS model infers the trend in the rate of speciation across the entire phylogeny through time (Höhna, 2015; Höhna et al., 2016; May et al., 2016).

To test our two hypotheses about primate biodiversity, we first compared the number of species identified in our phylogeny from each clade to the crown age we recovered for that clade and identified a linear relationship (R ² = 0.56), suggesting a correlation between age and species richness. Next, we extracted the mean empirical hyperparameter of the speciation rate from the TESS result for each clade and performed the same regression with species richness. However, we observed a much weaker correlation (R ² = 0.10).

The TimeTree database uses the NCBI taxonomy resource (Schoch et al., 2020) for its taxonomic framework, and the same has been applied for primates. We identified 462 binomial primate taxa recognized as valid in this reference, excluding extinct species (such as Homo heidelbergensis), species which could not be identified (often indicated with an “sp.” in place of a specific name), and any hybrids, redundant subspecies, or regional variants. This list formed the basis of subsequent literature searches to identify potential timetrees for our supertree approach. Of these, Cheracebus medemi, Callicebus oenanthe, Cheirogaleus andysabini, Cheirogaleus grovesi, Cheirogaleus minusculus, Cheirogaleus shethi, and Hapalemur gilberti have no molecular data deposited in GenBank which had been annotated with individual genes and had appeared in no molecular phylogenies we could identify, rendering them incompatible with the synthetic tree building approach we used here. Thus, our target species list included 455 primates (98% of the 462 species included in NCBI). These 455 species represent 87% of the 525 primate species recognized by the IUCN Red List (IUCN, 2024). The additional 63 species missing from NCBI lack molecular data.

Phylogenies were reproduced as Newick strings from published image files for seven studies first by manually constructing the topology in the alpha release of MEGA version 12 (Tamura et al., 2021). This new feature allows users to add, remove, and reposition phylogenetic branches in a graphical user interface and then export the result as a Newick string readable by any standard phylogenetic software. Branch lengths were measured using ImageJ 1.53k (Schneider et al., 2012) by recording the length of the provided scale bar in pixels and then translating the length of each phylogenetic branch in pixels into the provided units, either millions of years for timed trees or molecular substitutions for untimed trees.

We then timed the untimed trees using a literature consensus secondary calibration approach developed in previous work (Craig et al., 2023a; Craig et al., 2023b). For each of the five untimed trees, we constructed a relative timetree using RelTime in MEGA (Tamura et al., 2012; Tamura et al., 2018; Tamura et al., 2021). We then selected a relatively basal divergence, but not the crown split, as RelTime estimated divergence times for the ingroup species and treated this as a time calibration point. We used the TimeTree database to generate a distribution of divergence times estimated in prior published work. We assumed the minimum and maximum boundaries of the confidence interval around the median estimated time as endpoints of a uniform distribution imposed on the selected node. Using this calibration, we finally converted our relative timetree into an absolute timetree.

For the four species with molecular data existed yet no published phylogeny, we searched GenBank for a mitochondrial protein greater than 100 amino acids in length, which had been the focus of substantial prior research. This included CYTB and three subunits of NADH dehydrogenase.

We submitted these to NIH CGR SmartBLAST, and from the resulting accessions, we selected the accession with the highest percentage of shared identity for each primate species, plus the homologous human and mouse accession. We exported these to a fasta using GenPept, aligned and trimmed excess sequences from the ends where necessary, and built timetrees in MEGA (Tamura et al., 2021). We used the maximum likelihood search for each under the JTT model. We used little bootstraps (Sharma and Kumar, 2021) with an adaptive parameter search as a test of confidence in our topology. We then timed these trees using RelTime (Tamura et al., 2012; Tamura et al., 2018), providing both our alignment and the inferred ML tree, and calibrating 2–4 nodes per tree. We selected literature–consensus secondary calibrations from TimeTree as above for each genus, which was recovered as monophyletic or in the cases of genera for which fewer than half of the species were present, meaning that the deepest divergence we observed was likely not the true phylogenetic crown of the genus; we calibrated the divergence between the genus and its sister genus (the divergence between Nycticebus and Loris was calibrated this way). All alignments, calibration schemes, and intermediary tree files are available in the Supplementary Material.

The cladogenetic diversification rate shift (ClaDS) model (Maliet et al., 2019; Maliet and Morlon, 2021) infers the rate of speciation for each daughter lineage of a given phylogenetic divergence based on the species richness of the descendant clade. We ran the ClaDS model in Julia 1.9.3 (Bezanson et al., 2017), using an automatic cutoff for the Bayesian process at a convergence among three concurrent Markov chains when the Gelman statistic decreased below 1.05 (Maliet and Morlon, 2021). We imposed a sampling fraction of 1.0 as we present a near-complete phylogeny.

TESS (Höhna, 2015; Höhna et al., 2016; May et al., 2016; Fabreti and Höhna, 2022) estimates phylogeny-wide speciation through time as well as an initial hyperparameter of speciation for the whole clade. We allowed TESS to infer hyperparameters directly from each given phylogeny before each run and then ran the chain for 200,000 iterations, taking the first 10,000 as burn-in. We chose not to parameterize any mass extinctions. As in ClaDS, we imposed a sampling fraction of 1.0.

To combine all timetrees, we ran Chrono-STA (Barba-Montoya et al., 2024) using its default parameters. We used the first release of Chrono-STA, which is publicly available from its GitHub repository: https://github.com/josebarbamontoya/chrono-sta.