1. For taking a selfie to put on Insta, no, not really, a reproduction doesn’t really affect the outcome. It could even help things if there are parallel lines for selfies with an array of duplicate Mona Lisas
2. For appreciating the details of the composition and coloring, it has to be a fairly high quality replacement
3. To appreciate the brushstrokes, an even higher fidelity reproduction is required
4. To look at chemical composition of the pigments to understand what plants and minerals were used, there is no substitute for the actual physical object

I recently toured the dinosaur exhibit at the Natural History Museum London. It’s justifiably famous due to the museum’s important place in the history of paleontology. However, many of the specimens on display are fake.² There’s no trickery involved: in small lettering on the displays it lists that the dinosaur is a cast. In paleontology this is widely accepted. The actual fossils can be too heavy to mount without the risk of damage, and they can be hard to study in depth once mounted 15 feet in the air, so it’s common practice to make a model of the fossil and mount that. Sometimes these are cast directly from the original specimen; sometimes they’re cast from a previously made cast. Some museums are very direct about this – the Field Museum, for example, with its famous “Sue” T. rex skeleton has the actual skull fossil on display and signs explaining that the one on the full skeleton is a re-creation (and not just a cast). Some museums now also have the recreations or models in a different color from the actual fossils so visitors can see the difference, or a placard showing which bones are real fossils versus reconstructions. The Natural History Museum London’s dinoaur gallery seems to be largely an older exhibit (i.e., prominent signage indicating that dinosaurs, which are “reptiles”, went extinct at the end of the Cretaceous but birds, which are a different group (?!), did not) did not make any major distinctions that I noticed. There is evidence that visitors prefer original specimens, not replicas (i.e., van Gerven et al. 2018, Carsten Connor & Perin, 2020), but that’s not the same as they learn nothing from a replica – it’s a widely accepted practice for very good reasons.

So, why doesn’t every science museum have every dinosaur in life-sized replica form? Well, one thing is cost. A replica of the “Stan” T. rex costs $150,000, plus shipping from the Black Hills Institute, but one can also get a full-sized replica (perhaps less accurate) from a less well-known supplier for $50,000 and from another as cheap as $11,500. Space also matters – a herd of sauropods takes a lot more room, and brings in less money, than a second dino gift shop.

It does still surprise me a bit, though: taking the pricey but accurate options, one could get full adult replicas of theropods T. rex ($150K), Acrocanthosaurus ($172K), Allosaurus ($86K), and Struthiomimus ($65K), plus a Velociraptor and Protoceratops locked together in death ($13K), plus full adult replicas of Psittacosaurus ($9K), Triceratops ($108K), Stegosaurus ($94K), and Edmontosaurus ($90K), for only $786,000 from the Black Hills Institute, and perhaps as little as a tenth of that from some of the less famous companies. Moreover, this all seems to predate the revolutions in 3D scanning (even your phone can generate 3D models, let alone all the tech available with lidar and similar) and 3D printing that could make it even more accessible. The Metropolitan Museum of Art, for example, has high definition 3D scans of many pieces in its collection, and there are initiatives to share thousands of scans from museums for printing.

@online{o'meara2026,
  author = {O’Meara, Brian},
  title = {Museum {Shards}},
  date = {2026-07-19},
  url = {https://brianomeara.info/posts/museumshards/},
  langid = {en}
}

The poster is on display on Tuesday, June 23, 6 - 8:30 pm, Grand ballroom BC, Poster session 2: “Age-rate scaling in evolution is largely artifactual” Brian O’Meara and Jeremy Beaulieu

@online{o'meara2026,
  author = {O’Meara, Brian},
  title = {Age Rate Scaling Poster},
  date = {2026-06-17},
  url = {https://brianomeara.info/posts/ageratescalingposter/},
  langid = {en}
}

• Sunday, June 21, 10:00 am, Room 25C, Evolutionary ecology 1: “Rare today, common tomorrow: Evolutionary transitions and non-equilibrium dynamics shape species rarity”. Alivia Nykto and Brian O’Meara
• Monday, June 22, 4:15 pm, Room 22, Macroevolution: “Are the phenotypic extremes ecological generalists? Testing Liem’s paradox in a macroevolutionary framework” Sam Borstein (lab alum) and Matt McGee
• Tuesday, June 23, 6 - 8:30 pm, Grand ballroom BC, Poster session 2: “Age-rate scaling in evolution is largely artifactual” Brian O’Meara and Jeremy Beaulieu
• Wednesday, June 24, 2:30 pm, Room 21, Phylogenetic theory and methods: “Into the treeverse: Exploring the effects of topological uncertainty on comparative methods” Remington Motte, Brian O’Meara, Jeremy Beaulieu

There are a lot of great mixers, plenaries, and other events to attend. One thing I think folks overlook is the members’ meetings for the three societies. Do you think organizations of scientists should be speaking out on a particular issue? Directing funding towards students or postdocs? Improving a meeting in a particular way? Doing something new with a journal? Even changing the society’s name (something that was floated at an SSB meeting and later went out for a vote)? The members’ meeting is a good place to bring this up. Our societies have hundreds to thousands of members, run popular meetings, sponsor journals, and more, but a single voice in a members’ meeting can steer them in helpful new ways. The members’ meetings are also big enough that if you want to lurk at the back and just get the vibe, that’s ok, too.

Some other general meeting advice:

• When you stand in a group to talk, form a C shape, so there’s a hole for a new person to join. Once they join, open it up again.
• There’s an active code of conduct for the meeting; it applies to the attendees throughout the conference: not just at the conference center, but also, say, at the bar after sessions end for the day. See more here and do what you can to help keep the meeting safe for all. Our safety officer seems good. There are also EvoAllies and members of the Code of Conduct committee who can help link to resources, but the only one who can do investigations is the safety officer (she’s trained in how to properly interview respondents).
• Rooms are often crowded. I still try to protect others and myself by wearing a mask. Evolution no longer enforces masking, but anyone can choose to mask.

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2026,
  author = {O’Meara, Brian},
  title = {Evolution 2026 in Person Activities},
  date = {2026-06-15},
  url = {https://brianomeara.info/posts/evol2026inperson/},
  langid = {en}
}

In the US, a lot of interest focuses on proposed laws at the national level, but individuals’ lives can be dramatically affected by legislation passed at the state level. This post describes how I try to keep informed about legislation in the state I live in (Tennessee) but I suspect it will work well for other states, as well. I noticed that friends often are surprised by bills that come through, so I thought it would be useful to have a “how to” on staying aware of legislation.

As an example, let’s focus on House Bill 2194, Senate Bill 2259, from Tennessee’s 114th general assembly, as it will be of interest to others in academia. It was signed into law on April 16, 2026. Its short description, based on the bill at the time it was first introduced (known as caption text), is “Education, Higher - As introduced, requires the board of regents, state university boards, and the board of trustees for the University of Tennessee system to adopt and implement policies that clearly distinguish between tenure decisions and disciplinary actions for faculty members.” A couple of sections (but it is worth reading the full bill)

Prior to termination or suspension based on an allegation of misconduct, a tenured or non-tenured faculty member is only entitled to a written notice of the grounds for termination or suspension and an opportunity to be heard by the institution’s chief academic officer or chief executive officer. All terminations and suspensions based on an allegation of misconduct by the tenured or non-tenured faculty must be made by the institution’s chief executive officer or chief academic officer without any recommendation or vote by another faculty member at the institution.

and

The policies adopted pursuant to subsection (b) must: (1) Ensure that awarding, denial, or revocation of tenure is not used as a form of discipline; (2) Ensure that disciplinary actions do not alter or suspend a faculty member’s tenure status except as provided by institutional policy and after providing the faculty member due process; (3) Provide disciplinary procedures that are the same for tenured and non-tenured faculty for a faculty member’s misconduct; (4) Ensure due process for a faculty member; and (5) Comply with applicable state and federal law.

The point of this post is not to argue pros or cons of this bill (which has now been signed into a law, and takes effect on July 1, 2026), but rather for explaining how I notice and watch bills like this for those curious about future bills on related topics or other topics of interest can track their progress.

@online{o'meara2026,
  author = {O’Meara, Brian},
  title = {Reading Bills},
  date = {2026-04-25},
  url = {https://brianomeara.info/posts/readingbills/},
  langid = {en}
}

In this age of genomics, who cares about low sample sizes? Well, some charismatic clades folks study are naturally small (there are only so many baleen whale species – maybe you should have worked on insects instead?). But a more common issue will occur in approaches that try to deal with the differences across the tree of life by applying different rates to different parts of the tree. Older approaches like Medusa (Alfaro et al. 2009) and BAMM (Rabosky, 2014), and newer approaches like the Bayesian ClaDS (Maliet et al. 2019) or non-Bayesian MiSSE (Vasconcelos et al. 2022) also effectively subdivide the tree into different chunks (though this is done in a more continuous way with MiSSE, as it uses hidden states). So it’s possible a chunk is small enough that it has few enough taxa that it would be affected by bias in an estimator.

Jeremy Beaulieu and I have been on a mission lately to address bias in estimators (note this is a different issue than some of the myriad other issues affecting diversification analyses; see, for example, one of our preprints). There can be bias from ignoring measurement error (Beaulieu & O’Meara (2025); O’Meara & Beaulieu (2024); see also my upcoming poster at Evolution 2026), there can be bias from only looking at “interesting clades” (Beaulieu & O’Meara (2018 & 2019)) etc. Today we published the latest paper in this series (Beaulieu & O’Meara 2026).

Tanja Stadler (Stadler, 2013) wrote a key paper on estimators in diversification models, pointing out that the way different implementations estimate “the” speciation rate can vary and comparing different approaches. Inspired by this and some of our (well, largely Jeremy’s) work on finding an unbiased estimator in the Yule (pure birth, no extinction) case, we set out to find unbiased estimators for various diversification parameters.

The way we did this was rather fun. There’s the “real” math way of doing it by deriving equations – we were able to do this for Yule and the critical branching process (speciation = extinction), and for incorporating the effect of not looking at trees of two taxa or fewer. However, for the more complicated case where speciation and extinction can both vary, we could not analytically find a solution.

The approach we used instead has the polite name “symbolic regression” and is formally a machine learning approach (so maybe it aligns with NSF priorities for AI – program officers take note). But it boils down to, “well, we know a bias correction probably uses something about the number of taxa, and/or the age of the tree, and/or the speciation rate estimate, and/or the extinction rate estimate, and/or…” and “the correction might involve the likelihood estimate being multiplied by one (or more) of these, added to by one of these,…” and then trying ALL the possible equations to find the correction that works best. This follows my long-term strategy of “math can be difficult, let’s just make the computer work very hard” like the approach I used for parametric species delimitation (store the probabilities of all the gene trees (O’Meara 2009; see this header file)) or the phrapl series of papers for looking at phylogeography, gene flow, and species delimitation that involve estimating the probability of a gene tree by simulating many, many times. Here, we basically make all the reasonable corrections and try each to see which reduces the bias in the estimator. The key difficulty, as with approximate Bayesian computation, is to figure out a good way to estimate the distance from what we are getting and what we want. Something like RMSE is the obvious choice, but the issue is that for really tiny trees, the errors can be huge in a handful of simulations and those can overwhelmingly drive what is chosen. So instead we did a regression and saw how far the regression slope and intercept were from the ideal line.

We validated this approach (which used a slightly modifed gramEvol R package (Noorian et al. 2016) to search exhaustively) on the Yule case where we knew the correct result, then applied it to the more complex birth-death result.

Overall, it looked like the best unbiased estimator for speciation rate was the maximum likelihood estimate (MLE) for speciation rate times (number of taxa on the tree - 1) / (number of taxa on the tree - 2). The correction approaches 1 as the number of taxa on the tree (ntax) increases, but at just three taxa, the unbiased speciation rate is twice that of the MLE. The extinction rate, which is famously hard to estimate well, is the MLE of extinction rate times (ntax/(ntax-1) + extinction fraction), where extinction fraction is the MLE for extinction rate divided by the MLE for the speciation rate. However, read the paper for details.

It’s important to note that this is NOT another example of “diversification methods are wrong” – those that were wrong before are still wrong, and those that have yet to be proven wrong are still yet to be proven wrong¹. However, it does suggest that if your question is about the parameter estimates themselves, it could be worth converting them to the unbiased estimates, especially when comparing between groups of different sizes. I suggest phylogenetic software start returning both the MLEs and the unbiased estimates to users; users could also convert the estimates themselves, as these are basic transformations.

Another note: while we all like to gripe about peer review (somehow especially popular for those whose papers I review), this was a case where the manuscript was improved substantially by the people who volunteered to review it and AE Mike Steel, especially by one reviewer who caught an error in an earlier draft. Our thanks to the reviewers were heartfelt for this paper. This was my first time publishing in Bulletin of Mathematical Biology, which is published by Springer for the Society for Mathematical Biology and it was a delightful process – fast, competent reviews and a quick turnaround for publication.

@online{o'meara2026,
  author = {O’Meara, Brian},
  title = {Bias Correction in Diversification Models},
  date = {2026-04-17},
  url = {https://brianomeara.info/posts/biascorrection/},
  langid = {en}
}

Detailed instructions on using it and installing it (it’s large, so install_github() won’t work) are at https://bomeara.github.io/rnsf/index.html. It includes all the data cached plus various utility functions.

library(rnsf)
library(tidyverse)

── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr     1.2.0     ✔ readr     2.1.6
✔ forcats   1.0.1     ✔ stringr   1.6.0
✔ ggplot2   4.0.2     ✔ tibble    3.3.1
✔ lubridate 1.9.5     ✔ tidyr     1.3.2
✔ purrr     1.2.1     
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

library(DT)

data(grants)
grants$academic_semester <- rnsf::date_to_academic_semester(as.Date(grants$date, format="%m/%d/%Y"))
grants_aggregated <- grants |> filter(academic_semester %in% apply(expand.grid(c(2024:2026), c(" Fall", "  Spring")), 1, paste0, collapse="")) |> group_by(academic_semester, awardeeStateCode) |> summarise(total_awarded = n(), .groups = "drop_last") |> ungroup() |> tidyr::pivot_wider(names_from = academic_semester, values_from=total_awarded, values_fill=0) |> dplyr::arrange(desc(`2024 Fall`))
colnames(grants_aggregated) <- gsub("  ", " ", colnames(grants_aggregated))
colnames(grants_aggregated)[1] <- "Area"
grants_aggregated[,1] <- rnsf::abbreviation_to_state(unname(unlist(grants_aggregated[,1])))
grants_aggregated |> 
  DT::datatable(
    caption = "Funding by state or territory",
    filter="top") 

@online{o'meara2026,
  author = {O’Meara, Brian},
  title = {Rnsf},
  date = {2026-04-04},
  url = {https://brianomeara.info/posts/rnsf/},
  langid = {en}
}

• What was the discovery or innovation in this paper?
• How this was supported by evidence in the paper?
• What didn’t you understand about the paper, the methods, the questions, and/or the organisms?

I would use this info to tune what I covered in the intro for each class session. People brought a range of experience to the class, and I wanted to structure it so that everyone would learn a bit within each class session. Below are links to the posts for each week’s reading (slides, reflections on the paper, etc.):

• Tree building (Aug 22, 2025): Alison R. Irwin, Nicholas W. Roberts, Ellen E. Strong, Yasunori Kano, Daniel I. Speiser, Elizabeth M. Harper, and Suzanne T. Williams. 2025. “Evolution of Large Eyes in Stromboidea (Gastropoda): Impact of Photic Environment and Life History Traits” Systematic Biology 74(2):301–322. https://doi.org/10.1093/sysbio/syae063
• Bayesian analysis & fossilized birth death models (Aug 29, 2025): Laura P. A. Mulvey, Mark C. Nikolic, Bethany J. Allen, Tracy A. Heath and Rachel C. M. Warnock. 2025. “From fossils to phylogenies: exploring the integration of paleontological data into Bayesian phylogenetic inference” Paleobiology 51, 214–236. https://doi.org/10.1017/pab.2024.47
• Growth of the tree of life (Sep 05, 2025): Molly Chen, Artem I. Kholodov and Laura A. Hug (2025). “The evolution of the tree of life” Phil. Trans. R. Soc. B 380: 20240091. https://doi.org/10.1098/rstb.2024.0091
• Genomic species delimitation (Sep 12, 2025): Sonal Singhal, Adam D. Leaché, Matthew K. Fujita, Carlos Daniel Cadena, and Felipe Zapata. 2025. “A Genomic Perspective on Species Delimitation” Annu. Rev. Ecol. Evol. Syst. 2025. 56:467–89 https://doi.org/10.1146/annurev-ecolsys-102723-055311
• Phylogeography (Sep 17, 2025): Manica Balant¹, Daniel Vitales¹, Zhiqiang Wang¹, Zoltán Barina, Lin Fu, Tiangang Gao, Teresa Garnatje, Airy Gras, Muhammad Qasim Hayat, Marine Oganesian, Jaume Pellicer, Seyed A. Salami, Alexey P. Seregin, Nina Stepanyan-Gandilyan, Nusrat Sultana, Shagdar Tsooj, Magsar Urgamal, Joan Vallès, Robin van Velzen, Lisa Pokorny. 2025. “Integrating target capture with whole genome sequencing of recent and natural history collections to explain the phylogeography of wild-growing and cultivated Cannabis” Plants People Planet. 1-18. https://doi.org/10.1002/ppp3.70043 [¹ = equal contributions]
• Simulation (Sep 24, 2025): Ornela N. Dehayem, Ryan F. A. Brewer, Luis Valente, Frederic Lens, Rampal S. Etienne. 2025. “Impact of sampling strategy on inference of community assembly processes in phylogenetic island biogeography” Methods in Ecology and Evolution. 16:1507–1520. https://doi.org/10.1111/2041-210X.70058
• Behavior and genomics (Oct 3, 2025): Sara E. Lipshutz, Mark S. Hibbins, Alexandra B. Bentz, Aaron M. Buechlein, Tara A. Empson, Elizabeth M. George, Mark E. Hauber, Douglas B. Rusch, Wendy M. Schelsky, Quinn K. Thomas, Samuel J. Torneo, Abbigail M. Turner, Sarah E. Wolf, Mary J. Woodruff, Matthew W. Hahn & Kimberly A. Rosvall. 2025. “Repeated behavioural evolution is associated with convergence of gene expression in cavity-nesting songbirds” Nature Ecology & Evolution. https://doi.org/10.1038/s41559-025-02675-x
• Reticulate evolution (Oct 10, 2025): Gil Yardeni, Michael H. J. Barfuss, Walter Till, Matthew R. Thornton, Clara Groot Crego, Christian Lexer, Thibault Leroy and Ovidiu Paun. 2025. “The Explosive Radiation of the Neotropical Tillandsia Subgenus Tillandsia (Bromeliaceae) Has Been Accompanied by Pervasive Hybridization” Systematic Biology https://doi.org/10.1093/sysbio/syaf039
• Trait evolution (Oct 17, 2025): Verónica A. Rincón-Rubio, Rosana Zenil-Ferguson, Alejandro Gonzalez-Voyer. 2025. “The macroevolutionary consequences of the association between frugivory and carotenoid-dependent plumage coloration in passerine birds” Evolution, 2025, 79(8), 1643–1657 https://doi.org/10.1093/evolut/qpaf105
• Handling data deficiency in SDMs (Oct 24, 2025): Shubhi Sharma, Kevin Winner, Laura J. Pollock, James T. Thorson, Jussi Mäkinen, Cory Merow, Eric J. Pedersen, Kalkidan F. Chefira, Julia M. Portmann, Fabiola Iannarilli, Sara Beery, Riccardo De Lutio, Walter Jetz, 2025. “No species left behind: borrowing strength to map data-deficient species” Trends in Ecology & Evolution 40, 699–711. https://doi.org/10.1016/j.tree.2025.04.010
• Machine learning for traits (Oct 31, 2025): Roberta Hunt, José L. Reyes-Hernández, Josh Jenkins Shaw, Alexey Solodovnikov, Kim Steenstrup Pedersen. 2025. “Integrating Deep Learning Derived Morphological Traits and Molecular Data for Total-Evidence Phylogenetics.” Systematic Biology 74(3): 453-468 https://doi.org/10.1093/sysbio/syae072
• Alpha taxonomy and UCEs (Nov 7, 2025): Emma E. Jochim, James Starrett, Hanna R. Briggs, Jason E. Bond. 2025. “Speciation Pattern and Process in the California Coastal Dune Endemic Trapdoor Spider Aptostichus simus (Mygalomorphae: Euctenizidae) and Description of a New Cryptic Species” Ecology and Evolution 15:e72346 https://doi.org/10.1002/ece3.72346
• Multivariate traits (Nov 14, 2025): Emma Sherratt, Jenna Crowe-Riddell, Alessandro Palci, Ammresh, Mark N. Hutchinson, Michael S.Y. Lee, Kate L. Sanders. 2025. “Rapid evolution and cranial morphospace expansion during the terrestrial to marine transition in elapid snakes” Evolution 1–13 https://doi.org/10.1093/evolut/qpaf180
• Phylogenomics, recombination desert, and speciation (Nov 21, 2025): Nicole M. Foley, Richard G. Rasulis, Zoya Wani, Mayra N. Mendoza Cerna, Henrique V. Figueiró, Klaus Peter Koepfli, Terje Raudsepp & William J. Murphy. 2025. “An ancient recombination desert is a speciation supergene in placental mammals” Nature https://doi.org/10.1038/s41586-025-09740-2

Each week I’d decide on a topic, then look on OpenAlex, Google Scholar, and sometimes various journal websites for recent (only 2025) papers relevant to the topic. I would then read through these to find a paper that I thought would be useful for teaching, including whether I thought it was a good example of what to do (all papers have compromises, but I think students learn more from “good” papers than from ripping apart “bad” papers). I chose not to limit the papers only to ones in DAFNEE journals or only papers that are open access. There are issues with the publishing ecosystem, but for this class using that to limit papers would have resulted in too small a pool of papers this semester.

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Summary}},
  date = {2025-12-01},
  url = {https://brianomeara.info/posts/phylopapers_2025_summary/},
  langid = {en}
}

Nicole M. Foley, Richard G. Rasulis, Zoya Wani, Mayra N. Mendoza Cerna, Henrique V. Figueiró, Klaus Peter Koepfli, Terje Raudsepp & William J. Murphy. 2025. “An ancient recombination desert is a speciation supergene in placental mammals” Nature https://doi.org/10.1038/s41586-025-09740-2

This paper used a deep learning approach (from way back in 2020! Adrion, Galloway, & Kern 2020) to estimate recombination rate along multiple mammal species’ genomes. It also aligned chromosomes to each other to show how overall structure remained fairly stable across evolutionary time.

For teaching, this was useful at showing the utility of using sliding windows when scanning along a genome, deep learning for parameter estimation, and looking at conflicts between gene tree and species tree topologies and the frequencies of the different possible gene trees. This paper was also nice for showing the impact of very data-rich graphics that still communicate a clear message. The potential impact of genes on the X chromosome for speciation is also something worth investigating more.

One reason I was excited to teach a class like this was to help my learning, which happened every week. The fun aspect of that in this paper was the conclusion that to understand phylogeny in tricky situations (such as ongoing gene flow) it could be better to focus on the non-recombining core of the X chromosome. I remember the days when we all used just mtDNA, then maybe a few nuclear genes, now various ways to sample across the entire genome. Going back to effectively a single history is counterintuitive, but after reading this paper it makes sense. I could see a use case for development of recombination-rate-aware gene tree - species tree approaches: use many genes and accommodate the realities of different gene histories, but with weighting so that areas with lower recombination rates have a greater weight in the final reconstruction. One could do a fast, dirty approach where one assigns weight to different genes (such as genes on the X chromosome having w times the weight of others, and do a sensitivity analyis of w) but if recombination is inferred as part of the analysis the impact would flow in naturally from the model.

I made intro slides with some of my background material and some figures from the paper: PDF and PowerPoint.

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Phylogenomics}},
  date = {2025-11-21},
  url = {https://brianomeara.info/posts/phylopapers_2025_Nov_21/},
  langid = {en}
}

Emma Sherratt, Jenna Crowe-Riddell, Alessandro Palci, Ammresh, Mark N. Hutchinson, Michael S.Y. Lee, Kate L. Sanders. 2025. “Rapid evolution and cranial morphospace expansion during the terrestrial to marine transition in elapid snakes” Evolution, 2025, 0(0), 1–13 https://doi.org/10.1093/evolut/qpaf180.

The paper looks at skull evolution changes as snakes went back to the sea. It is a great, modern take on how geometric morphometrics works, using X-ray microcomputed tomography of skulls. It uses landmarks to measure shape and a variety of approaches to visualize changes and to estimate rates of change. It is also an opportunity to introduce students to the Procrustes myth (namesake of a statistical method).

One especially nice feature of this paper was its use of simulations – always a good way to get a sense of what the fitted models are actually saying about the morphology.

A necessary caveat when looking at spectacular evolutionary one-offs is that they are just that – N=1 examples. In this case, the group radiated into many species (and there’s also a different origin of a semi-aquatic species) but no matter how rigorous the statistics used are, it’s still a single origin of the change and the putatively correlated trait. The fully marine species do have a different shape and faster rate of evolution leading to them, but it could be due to something else changing on that branch besides the habitat shift. For example, if there were sexual selection for a sleeker head in that lineage, we would not know that led to the shift in shape rather than habitat – maybe the head shape shift unlocked the ability to be marine. This isn’t a criticism of this paper, but rather something everyone working on single synapomorphies must wrestle with, whether it’s the consequences of evolving a flower, vertebrate transition to land, or beaks in birds. It’s still possible to reject some hypotheses – for example, this paper had enough power that had it shown there was no association of the marine shift with head shape change, I would believe it, so showing there is an association present does give more weight to the idea that one change led to the other. It’s not the same as finding the same head shape change in five different cases when snakes went back to the ocean, but we just don’t have five such examples.

Though it wasn’t a focus of this paper, its mention of earlier work showing a decoupling of rates between speciation and morphology was also a benefit. We often have what I call the Martha Steward macroevolutionary hypothesis: “It’s a good thing” [the catchphrase of this American TV and cookbook personality]. We assume that some presumably adaptive trait will lead to increased speciation, lower extinction, and faster rates of evolution overall, even when there is no clear mechanism as to why (for example, perhaps living in the ocean results in fewer mating barriers and so should slow the rate of species formation). Work like that in this group shows that these rates can in fact operate independently of each other (and that rate shifts that are detected aren’t driven by confounding factors, like branches that are estimated to be too short in one clade that then results in too high rate estimates).

I made intro slides with some of my background material, especially about Ornstein-Uhlenbeck models, and some figures from the paper: PDF and PowerPoint. Sample figure from the slides, showing how OU models work (sigma is the amount of sugar the kid has been fed, alpha is the strength of pull of his cord, and theta is where he is being pulled to):

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Multivariate} Traits},
  date = {2025-11-14},
  url = {https://brianomeara.info/posts/phylopapers_2025_Nov_14/},
  langid = {en}
}

Emma E. Jochim, James Starrett, Hanna R. Briggs, Jason E. Bond. 2025. “Speciation Pattern and Process in the California Coastal Dune Endemic Trapdoor Spider Aptostichus simus (Mygalomorphae: Euctenizidae) and Description of a New Cryptic Species” Ecology and Evolution 15:e72346 https://doi.org/10.1002/ece3.72346.

This revisits a clade of trapdoor spiders in California. It is noteworthy that even in a place with as many resources and researchers as California, and with organisms this charismatic, there remains open questions in taxonomy.

This paper used ultraconserved elements, UCEs, which have been popular recently as they have a mix of rates, so in theory some are useful for the questions being studied (and they come from across the genome):

Students engaged well with this paper, asking good questions about sampling, amount of evidence for the species boundaries, and the various methods used.

One thing that really impressed me about this paper was that it included the actual alpha taxonomy. I develop species delimitation methods, and so many published uses are basically “look, there are undescribed species here” and then the paper stops. This paper formally describes the new species: diagnostic characters, types, the works. Species boundaries are hypotheses, and I could imagine ongoing discussions about whether this group is under- or oversplit, but actually going from model conclusions to updating the taxonomy is a great step.

The paper also spurred good discussions about the practice of naming in taxonomy. The new species is named after Dr. Martina G. Ramirez, who has worked on spiders for decades, and who sounds like an amazing mentor for her students, so this naming avoids many of the issues of naming after people in the past with problematic impacts (see my SSB presidential address on this and related issues). However, naming organisms after people is very much an open question in our field, with reasonable advocates on many sides, so it was useful for students to work through the arguments in class to help understand the practice of taxonomy better.

I made intro slides with some of my background material and some figures from the paper: PDF and PowerPoint.

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Alpha} Taxonomy},
  date = {2025-11-07},
  url = {https://brianomeara.info/posts/phylopapers_2025_Nov_7/},
  langid = {en}
}

Here we present Kosmos, an AI scientist that automates data-driven discovery. Given an open-ended objective and a dataset, Kosmos runs for up to 12 hours performing cycles of parallel data analysis, literature search, and hypothesis generation before synthesizing discoveries into scientific reports. Unlike prior systems, Kosmos uses a structured world model to share information between a data analysis agent and a literature search agent. The world model enables Kosmos to coherently pursue the specified objective over 200 agent rollouts, collectively executing an average of 42,000 lines of code and reading 1,500 papers per run. Kosmos cites all statements in its reports with code or primary literature, ensuring its reasoning is traceable. Independent scientists found 79.4% of statements in Kosmos reports to be accurate, and collaborators reported that a single 20-cycle Kosmos run performed the equivalent of 6 months of their own research time on average.

I’m not planning to use this (my accuracy is better than a C+, I’d hope), but I imagine colleagues will, and this is just its first generation. And researchers and others are using ChatGPT and related approaches to help with papers already, despite the known issues with made up citations. Popular app Grammarly now offers a service that lets users write scientific statements (their example is from geology) and then it automatically finds and cites relevant papers – no need to actually read the literature, just decorate your text with the funny names and numbers in parentheses that make it look credible. Grammarly claims it is both a “partner for better grades” and will help you “turn claims into supported arguments.”

There’s a whole discussion to have about the ethics of using such services and the wisdom of the many colleges paying for their communities to have access to them. But I want to talk about something different: which papers these services will choose.

Search Engine Optimization (SEO) has long been a concern for website owners: how do we make it so that when people search “vacation in Aruba” my hotel site is the top hit? That matters for people in science, too (people on the job market, make sure you have a good website!), but it’s less of an emphasis. For writing papers, there’s some optimization (like the endless debate over the wisdom of including colons in titles), but it’s a messy, human process. And journals have various tactics for bumping up their impact factor (“Hey, want to publish version 1.2.1 of your popular software package in our journal?”; “Has there been a really good review of this popular subject published this week? Let’s commission one and get those citations!”). But increasingly, the entities “reading” our papers and deciding what to cite won’t be humans: they’ll be AI recommendation engines telling people what is most relevant. Even once the rumored AI economic bubble bursts, I suspect it’s unlikely tools like this will go away, unless there becomes enough of a pushback on their use that it becomes unprofitable to provide them.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {AI Optimization},
  date = {2025-11-07},
  url = {https://brianomeara.info/posts/ai_optimization/},
  langid = {en}
}

Roberta Hunt, José L. Reyes-Hernández, Josh Jenkins Shaw, Alexey Solodovnikov, Kim Steenstrup Pedersen. 2025. “Integrating Deep Learning Derived Morphological Traits and Molecular Data for Total-Evidence Phylogenetics.” Systematic Biology 74(3): 453-468 https://doi.org/10.1093/sysbio/syae072

Morphological traits have a long history in phylogenetics, and for some questions, like for reconstructions of long-dead species, they can be the only data available. This paper takes advantage of deep learning and a set of reference images for beetles to identify traits to use for phylogenetic reconstruction.

One useful bit of background was this introduction to convolutional neural networks (original page; archived). Note that “AI” covers a whole set of technologies, from ways to get text-based responses trained from a large corpus of text (“cheatGPT, write a five paragraph essay on the use of imagery in The Scarlet Letter”) to generating pictures or videos, to matching images or sounds to species (as in iNaturalist or Merlin); convolutional neural networks (CNN) are common ways to extract features and other information from images.

I was expecting this paper to extract discrete characters, as that are what are used most commonly in phylogenetics for tree inference: things like presence or absence of “pygidium exposed” or “geniculate antennae.” Instead it used continuous traits. Those are commonly ratios or other concrete trait measurements: head width divided by length, angle between two elements, etc. This uses features that the machine learning approach discovered but which are harder to map back to traits humans identify; for example, here’s which pixels matter to one of the traits used:

This paper shows these traits are informative and useful for recovering the phylogeny. It does not shy away from potential disadvantages, either, including the work still required and even the environmental cost of running these models.

Another thing this paper does excellently is providing supplementary data:

The data underlying this article are available at <http://doi.org/10.17894/ucph.39619bba‐4569‐4415‐9f25‐d6a0ff6 4f0e3> for the Rove‐Tree‐11 dataset and in the article’s dryad repository (<https://doi.org/10.5061/dryad. 9cnp5hqqq>) for the further molecular data and associated genbank accession numbers, example inference code, all generated trees, and stratified dataset split. All trained model runs and extracted trait matrices are available in the following erda repository https://erda.ku.dk/archives/440063cabdb1789ad82f31366c926b4e/published‐archive.html. The reference tree, best molecular tree and best total‐evidence tree can be found on TreeBASE at http://purl.org/phylo/treebase/phylows/study/TB2:S31300?x‐access‐code=397cc12bd8047bf52b312b4743f23e2b&format=html. The code used in this analysis is available on github https://github.com/robertahunt/Revisiting_Deep_Metric_Learning_PyTorch, commit a6654453c3b7785a17511255e02c468c53fe6f5d, forked from Roth et al. (2020).

It even includes putting the trees on TreeBase, something few in our field do despite the benefits to all (and citation bump for people who share).

I made intro slides with some of my background material and some figures from the paper: PDF and PowerPoint.

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Machine} Learning for Traits},
  date = {2025-10-31},
  url = {https://brianomeara.info/posts/phylopapers_2025_Oct_31/},
  langid = {en}
}

Shubhi Sharma, Kevin Winner, Laura J. Pollock, James T. Thorson, Jussi Mäkinen, Cory Merow, Eric J. Pedersen, Kalkidan F. Chefira, Julia M. Portmann, Fabiola Iannarilli, Sara Beery, Riccardo De Lutio, Walter Jetz, 2025. No species left behind: borrowing strength to map data-deficient species. Trends in Ecology & Evolution 40, 699–711. https://doi.org/10.1016/j.tree.2025.04.010

Figuring out where species can live now, where they could live in the future as climate shifts or habitat changes, and where they may have lived in the past are key questions for which species distribution models (SDM) can be helpful. There are many approaches, but in general they use many location points for a species, gather info on climate or other factors at those points, and use them to model where a species can and cannot occur. For example, palm trees do not do well where it is cold enough for them to freeze: all their points will be from areas without multiple freezing days, and a good model will use this and other factors to predict where they can live (and perhaps moving poleward as climate tends to warm). But the problem is that many species lack the data required to do this well.

This paper sketches out a few ways to use information from a related species to help model a focal species. These can include using phylogenies to connect predictions or parameters between species as well as various other ways, such as co-occurrence, to have data-rich species help make inferenes for data-poor species.

I like this sort of work because it makes continuous how we often treat individual species. We say we avoid typological thinking (“any member of this group is fundamentally the same as any other member”) but we de facto use this when we assume all members of a species are identical. The reality is that there will generally be more or less closely related subpopulations that are not 1:1 replacements for each other (though likely exchanging genes), but it’s reasonable that the populations are pretty good predictors for each other. However, for traditional approaches to SDM, we assumes this predictive similarity stops at the (somewhat arbitrary) “species” boundary. With some of the methods in this paper, some info can flow between related species. It is a good use for phylogeny, especially in a world where the amount of data can vary so dramatically between species. Another example of this was done by Jess Welch and Jeremy Beaulieu for predicting bat extinction risk.

Another advantage of this paper for teaching this week is that it was a relatively short review paper (TREE). It can be a nice break for students more used to dealing with methods-heavy empirical papers.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Handling} {Data} {Deficiency}},
  date = {2025-10-24},
  url = {https://brianomeara.info/posts/phylopapers_2025_Oct_24/},
  langid = {en}
}

Verónica A. Rincón-Rubio, Rosana Zenil-Ferguson, Alejandro Gonzalez-Voyer. 2025. “The macroevolutionary consequences of the association between frugivory and carotenoid-dependent plumage coloration in passerine birds” Evolution, 2025, 79(8), 1643–1657 https://doi.org/10.1093/evolut/qpaf105

This is an example of a thorough study that investigates trait evolution well. The paper looks at potential correlation between eating fruits and having feathers that use carotenoids for color. It’s fundamentally a trait correlation question, but the paper is careful to first check for possible association between the traits and diversification. As Wayne Maddison showed nearly two decades ago, even if your question is about traits, ignoring the effect of differential diversification can cause problems: there are more state 1 than 0 at the tips because of a faster 0->1 than 1<-0 rate, and/or higher diversification rate in state 1, and/or there hasn’t been enough time to move on from a state 1 ancestral state: one can’t just look at the transition rates alone. So this paper does a ton of work to examine potential confounding effects, finds none, and then goes on to find that the ancestor for passerines likely had carotenoid-dependent plumage and that in species that do not have both carotenoid-dependent plumage and frugivory they lose the other trait quickly as well.

If I were to have a quibble with the paper, it would be its lack of units: for example, the rate of loss of carotenoid pigmentation in birds that don’t have frugivory (Fig 2) is “0.06” – no units indicated in figure or caption. Omitting units is standard for our field, but it can help with interpretation. For example, 0.06, assuming the tree branch lengths are in units of millions of years, is 0.06 events/MY or, perhaps more intuitively, an expected wait time per species of 1/0.06 = 16.7 MY to lose frugivory. Putting rates in terms of expected time until a change can help show if numbers are reasonable: this one, for example, suggests a leisurely amount of time for the change but still evolutionarily feasible: it’s not like an expected time of 10 years nor of 10 billion years.

Overall, though, this is a great paper. Another teachable moment from it – it could have come across as a boring null result: “carotenoids for plumage (or frugivory) do not affect speciation or extinction” and maybe even left in a drawer. But it’s not! For one thing, the lack of correlation is a discovery – I would have thought that carotenoid-dependent plumage would increase speciation (by allowing more reinforcement of mating barriers that appear). And there is an additional story about the trait correlation themselves, and potential ancestral states, that help us understand evolution better.

I made intro slides with some of my background material (SSE models, trait models) and some figures from the paper: PDF and PowerPoint.

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Trait} Evolution},
  date = {2025-10-17},
  url = {https://brianomeara.info/posts/phylopapers_2025_Oct_17/},
  langid = {en}
}

Gil Yardeni, Michael H. J. Barfuss, Walter Till, Matthew R. Thornton, Clara Groot Crego, Christian Lexer, Thibault Leroy and Ovidiu Paun. 2025. “The Explosive Radiation of the Neotropical Tillandsia Subgenus Tillandsia (Bromeliaceae) Has Been Accompanied by Pervasive Hybridization” Systematic Biology https://doi.org/10.1093/sysbio/syaf039

This looks at subgenus Tillandsia inside genus Tillandsia, a group of morphologically quite diverse plants.

This paper was good for teaching because it was fairly focused: what is the phylogenetic network of the group and how is gene flow affecting this?

It’s yet another paper showing the utility of deleting massive amounts of data – in this case, looking at windows of the genome far enough apart that there is likely ample recombination between them. It was also useful in showing how one has to compromise. They used ASTRAL-III to infer a phylogenetic tree of all their samples incorporating processes like incomplete lineage sorting that might case gene tree - species tree mismatch. But they then used a much reduced number of taxa to analyze the history with zero to three possible reticulation events; given the massively larger search space when going from a tree to a network, they used a reduced set of samples. These were reasonable choices, but it shows the need for software that can feasibly look for reticulations for problems of hundreds of taxa or more.

The paper also used ABBA-BABA and related tests to scan along the genome to look for areas behaving unusually. Students have trouble understanding these at first (the ratio of the ‘wrong’ gene trees for sets of four lineages [ones that don’t match the species tree] having information about hybridization is a weird concept) but I think this paper explains it well, especially for an empirical paper not tasked with doing a review. It was also nice to see this being done as a genome scan as a discovery process. A lot of especially intro science is framed as rejection of dull hypotheses (“are these two different things not the same?”), but there’s still a lot of need for “let’s strap a camera on a sub and see what’s down there”, “let’s look at that galaxy in infrared”, “let’s see what Tillandsia are doing with their chromosomes”.

Side note I only briefly touched on in class: subgenus Tillandsia within Tillandsia?! “I study Tillandsia species,” even for someone using the same taxonomic concepts as oneself, is an ambiguous statement without including the rank. Yet another advantage of phylocode, where Tillandsia would have only one definition (and if one wanted to name smaller groups within it, one could without having to use subgenera, subsubgenera, superspecies, etc.). Unfortunately, Tillandsia as of today doesn’t seem to be defined at all in phylocode based on the online name repository; it might be in the Phylonyms companion volume to the code itself, but I lack a copy (but it’s on sale for $240 hardcover, only $225 for the eBook, so…).

I made intro slides with some of my background material and some figures from the paper: PDF and PowerPoint.

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Reticulate} Evolution},
  date = {2025-10-10},
  url = {https://brianomeara.info/posts/phylopapers_2025_Oct_10/},
  langid = {en}
}

Sara E. Lipshutz, Mark S. Hibbins, Alexandra B. Bentz, Aaron M. Buechlein, Tara A. Empson, Elizabeth M. George, Mark E. Hauber, Douglas B. Rusch, Wendy M. Schelsky, Quinn K. Thomas, Samuel J. Torneo, Abbigail M. Turner, Sarah E. Wolf, Mary J. Woodruff, Matthew W. Hahn & Kimberly A. Rosvall. 2025. “Repeated behavioural evolution is associated with convergence of gene expression in cavity-nesting songbirds” Nature Ecology & Evolution. https://doi.org/10.1038/s41559-025-02675-x

[Note my conflict of interest: I was on Sara Lipshutz’ PhD committee]

One of the students in the class requested a modern paper on behavior. This paper had a lot of material for students to parse, but I think it’s good for them to see what a modern, chewy paper looks like. It has everything from careful natural history observations of birds attacking models to modern genomics to measures of hormones.

In all my classes I emphasize the difficulty of discretizing biology. Nature is full of variation. Even things that seem easiy countable, like number of limbs, have had variation over time and within species (quick, how many femurs do snakes have? Two in pythons). For this paper, a question was about aggression in cavity-nesting birds (birds nesting in holes in trees, for example) versus ones making open nests (the classic “use a crayon to draw a bird nest” nests). But even there, some birds only use cavity nests, some only use open nests, and some use both – how to categorize those? And does it matter if 99% of pairs in a species use a cavity or 3% of pairs use a cavity? Even continuous measures, like amount of aggression towards a model armed with a bluetooth speaker, require careful decisions. Is glaring at the model included in time spent in aggression? Flying around it yelling? Making full contact? This paper was useful in illustrating how these decisions are made.

The paper also had an interesting study design for a phylogenetics class. The authors chose non-overlapping pairs of species that differ in nesting habit. This is good, as it’s a lot like doing a twin study, but there still is differential relatedness between the pairs and so the authors used phylogenetic linear mixed models (PGLMMs – note, the “LLM” is not large language model) to control for relatedness. They also poked at the data in other ways to see if there could be confounding effects – for example, perhaps depth of the split between species in the pair had an effect on the results.

The genomic aspects I think were also informative. They looked for groups of genes that were expressed consistently differently between cavity and open nesters. There were a few found, but it felt appropriately preliminary: it wasn’t presented as “these are the genes that lead to aggression” but rather as candidates for potential further study. It was also an opportunity to teach students about how we understand putative gene function (from gene ontology databases).

I made intro slides with some of my background material and some figures from the paper: PDF and PowerPoint.

[Note that this blog post is dated for the class date, but I’m actually pushing it on Oct. 11, 2025]

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Behavior} and Genomics},
  date = {2025-10-03},
  url = {https://brianomeara.info/posts/phylopapers_2025_Oct_03/},
  langid = {en}
}

Ornela N. Dehayem, Ryan F. A. Brewer, Luis Valente, Frederic Lens, Rampal S. Etienne. 2025. “Impact of sampling strategy on inference of community assembly processes in phylogenetic island biogeography”. Methods in Ecology and Evolution. 16:1507–1520. https://doi.org/10.1111/2041-210X.70058

Simulation is used a lot to validate methods, but I like teaching it to students because it’s also important in power analyses. In my experience it is still rare for students to figure out if their proposed research will have enough species, data, etc. to potentially answer their question before embarking on it. It can save so much future heartache to spend a few days seeing whether, for example, if this trait leads to evolution of this other trait at a particular rate there’s any chance of being confident in a result from a likely 17-taxon tree.

However, simulation is hard to do well. For one thing, the space of parameters to examine can grow very quickly. One of the reasons I first learned R was because I had done a bunch of simulations (perl to test a C++ program) to test some methods I made for species delimitation (O’Meara 2010) but I hadn’t realized how hard it would be to plot something like 6 variables at once (so why not bar charts on top of bar charts in an array of bar charts?). “Number of taxa might matter… and speciation rate… and the rate of trait evolution… and tree shape… and age… and…”. It can also be hard to figure out for a given parameter what values to try. Does this new method to estimate different substitution rates work well? On a three-taxon tree with mis-estimated branches, nope, it works horribly. On a 10,000-taxon tree where branch lengths are perfectly correlated with time, it works splendidly. But what will be relevant to biologists? That’s one of the reasons Jeremy Beaulieu and I (Beaulieu & O’Meara 2015) made sure to include units in a simulation so people could see whether they’re reasonable (even though one reviewer asked for us to delete units!). The ability to pre-determine the outcome of a simulation can also lead to odd choices depending on who is doing the simulating. For example, one paper in our field evaluated the performance of a Bayesian method written by many of the same authors by using simulation parameters that match the priors used in the later analysis. Unsurprisingly, it worked well when its prior was centered on the truth – with such a sim, it might work even better without any data! On the other hand, one could simulate using models that violate the assumptions of a method and then show that the method fails (“the normal distribution is a terrible way to get confidence interval for a mean… when data are simulated from a univariate uniform distribution”).

Dehayem et al. (2025) is interesting because it’s a paper testing a method created by many of the same authors, so presumably they’d be more accepting of results that show their method works, but it does a careful job. It is handling a complex scenario: arrival and diversification on islands, including hard to estimate parameters like extinction rate. There are some needed assumptions made in the simulation (for example, that this model describes the process so the parameter values are meaningful). It gets parameters from previously fit biological datasets, which is a good thing to do as it centers them on presumably realistic values. They also tried various ways to violate the model assumptions, such as a bias against sampling young species (there may be no gene flow, but humans haven’t recognized the populations as different species yet). There are many other potential things to vary, but it keeps it pretty focused and thus understandable. This made it an accessible jumping off point for discussions of simulations.

I made intro slides with some of my background material and some figures from the paper: PDF and PowerPoint.

[Note that this blog post is dated for the class date, but I’m actually pushing it on Oct. 11, 2025]

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Simulation}},
  date = {2025-09-24},
  url = {https://brianomeara.info/posts/phylopapers_2025_Sep_24/},
  langid = {en}
}

Manica Balant¹, Daniel Vitales¹, Zhiqiang Wang¹, Zoltán Barina, Lin Fu, Tiangang Gao, Teresa Garnatje, Airy Gras, Muhammad Qasim Hayat, Marine Oganesian, Jaume Pellicer, Seyed A. Salami, Alexey P. Seregin, Nina Stepanyan-Gandilyan, Nusrat Sultana, Shagdar Tsooj, Magsar Urgamal, Joan Vallès, Robin van Velzen, Lisa Pokorny. 2025. “Integrating target capture with whole genome sequencing of recent and natural history collections to explain the phylogeography of wild-growing and cultivated Cannabis”. Plants People Planet. 1-18. https://doi.org/10.1002/ppp3.70043 [¹ = equal contributions]

This study examines wild and cultvated Cannabis, which is according to the paper has been used “as fibre (ropes, fabric and paper), medicinally (over 200 recorded uses), as food (nutrient-rich seeds) and in various magico-religious rituals”. It uses many samples from around the globe and genomic data to examine population structure.

One thing that has stood out in this paper and others considered for the course is the now apparent standard toolset of using mafft, Structure, ASTRAL-III, and IQ-TREE2. The age of some of the software is unexpected: the latest release of Structure is from over 13 years ago, and for ASTRAL-III it’s over five years ago. That’s not necessarily a bad thing – it’s not like hackers are trying to take over the banking system by exploiting potential vulnerabilities in unpatched phylogenetic software – it’s just surprising. You would think that especially for popular analytical questions there would be new innovations or just new hotness. For ASTRAL the developers “encourage using the new code” of ASTER, though as this only was published in a peer-reviewed journal in July 2025 it’s reasonable that papers we’re reading now don’t use it yet (and I doubt answers under ASTRAL-III are wrong). But I do worry as researchers continue to use ChatGPT and similar for analyses and ask “what’s the problem in doing stats with an (AI) consultant?” popular approaches will be baked into the training data and then will be continually suggested as the options to use, even when papers come out showing problems with classic approaches (and see all the studies STILL looking at net diversification rate through time) or new approaches unlock new questions. There’s already a bias towards that built into the field – it’s more efficient to use approaches one already knows (and you might know the limitations well, too, which is important), and all the tutorials or posts are about the classic software, but at least new info is constantly being put into our human brains, where we could have a preference towards recent approaches (though on R-sig-phylo there was a counterargument that the agentic AIs will learn about new approaches and suggest them). Though I guess one advantage of fossilization of methods advice is we’ll all start using MacClade again.

One thing I loved about the paper was the use of herbarium and fresh specimens. It’s yet another example of the benefits of repositories of biological information (recognized at many, but not all, places) as well as expertise to collect in the field. The paper also demonstrated the feasibility of genomic scale data.

A shift in thinking I’ve undergone but still feels unnatural is how in modern evolutionary biology a key workflow step is massively deleting data. It wasn’t that long ago we were looking at chromatograms in Sequencher to identify every possible base by eye (while still excluding bad reads); in contrast, in this paper the data were filtered from 68,212 single nucleotide polymorphisms (SNPS) down to just 2,875. This is important to do for data from GBIF (“nope, that oak is not from the ocean, someone flipped a sign”) and things like the TRY database of plant traits, too. This paper was useful for discussing this shift with students.

A conclusion from the paper was that Cannibis sativa is one species (as hypothesized by Linnaeus) and not two with the addition of C. indica. Perhaps a useful anecdote for intro bio when the standard move is to dunk on “Lamarckism” (though this requires more nuance than the cartoon version): the describer of the incorrectly split C. indica: Jean-Baptiste Lamarck. Though one student raised a good point: it would be interesting to redo the Structure analyses using only the wild plants, not escaped or domesticated cultivars – one possibility is that human meddling allowed interbreeding of C. sativa and C. indica that would have been relatively reproductively isolated otherwise.

One note for those who might want to teach the paper: I made sure to let students know that some uses of the plant are illegal in this jurisdiction (true at both our state and federal level) before asking them to provide information (for discussion questions on the paper before class, for example). I don’t want students trying to, say, make a joke, put in writing something that could be read by others as disclosure of illegal drug use – who knows how such info could be scraped into systems and misinterpreted in the future.

I made intro slides with some of my background material and some figures from the paper: PDF and PowerPoint.

[Note that this blog post is dated for the class date, but I’m actually pushing it on Oct. 11, 2025]

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Phylogeography}},
  date = {2025-09-17},
  url = {https://brianomeara.info/posts/phylopapers_2025_Sep_17/},
  langid = {en}
}

Sonal Singhal, Adam D. Leaché, Matthew K. Fujita, Carlos Daniel Cadena, and Felipe Zapata. 2025. “A Genomic Perspective on Species Delimitation” Annu. Rev. Ecol. Evol. Syst. 2025. 56:467–89 https://doi.org/10.1146/annurev-ecolsys-102723-055311

In contrast to the other papers so far in the class, this is a review paper rather than original research (though, like many modern review papers, it actually does a fair amount of new bibliometric analysis). I’m not going to have us read a lot of review papers, but this one touched on various topics that came up in previous discussions (like the multispecies coalescent) and was a bit of a breather for students in that it had less jargon. They loved it and found it understandable. I think it was a very thorough and fair overview of the subject.

The two biggest questions students had were about the machine learning mentions and the gdi metric (Jackson et al. 2017). People now often think of AI as large language models like ChatGPT, but it covers a wide variety of approaches: image recognition, like that used by iNaturalist; approaches to impute missing data; methods to infer relationships from large sets of data; etc. One could imagine hoping for a genomic equivalent for the long wished-for “barcoding gap” (a discrete difference between intra- and interspecific differences), for example. The gdi is a measure we came up with to compare differences between potential species: it is scaled from 0 for panmixia to 1 for strong divergence. (I actually fought my coauthors, because I didn’t really like the gdi, but my collaborators won out and they were right). I think part of the confusion from the Singhal et al. paper is that the gdi in their Fig 2 is shown as a property of a population, when it’s actually a measure used between two populations.

One thing that both I and the attendees liked was the discussion at the end of the paper about not making genomic approaches a requirement for species delimitation publications. Genomics brings a lot of possibilities (such as looking for genomic regions correlated with lack of gene flow), and it may even be cheaper than older sorts of data which have less information. But genomic data are still not feasible for all researchers or locations. Especially given the need to do basic species discovery in so many groups, using available data to understand biodiversity is more critical than waiting until one can do multiple full genomes.

An interesting data point from the bibliometric analysis was that only in 36% of cases did discoveries about a need to change taxonomy lead to a taxonomic change in the papers. This is in line with my anecdotal experience. There are good reasons for this (a taxonomic change is a big deal and requires expertise – one doesn’t want to rush into it) but it is a bit odd that methods intended to discover new species or collapse oversplit ones usually don’t lead to this in a usable way.

Overall, I expect this paper to become widely used in teaching and for people getting into the field.

I made intro slides with some of my background material and some figures from the paper: PDF and PowerPoint.

To subscribe, go to https://brianomeara.info/blog.xml in an RSS reader.

@online{o'meara2025,
  author = {O’Meara, Brian},
  title = {PhyloPapers 2025, {Genomic} Species Delimitation},
  date = {2025-09-12},
  url = {https://brianomeara.info/posts/phylopapers_2025_Sep_12/},
  langid = {en}
}