Ana Serra Silva
@anaserrasilva.bsky.social
770 followers 570 following 62 posts
Phylogeneticist and evolutionary biologist. Research fellow at UCL Biosciences | Communications Officer for @systassn.bsky.social
Posts Media Videos Starter Packs
anaserrasilva.bsky.social
Quarterly reminder to not write posts during my commute/ before finishing my coffee... ☕

That should be 6PM Friday, OCT 24!

Do join us ON THE 24TH for the @systassn.bsky.social Founders' Lecture.

More details here: systass.org/founders-lec...
anaserrasilva.bsky.social
Please register or get your students to register!

I do not want to unleash nagging mode!
systassn.bsky.social
The Young Systematists' Forum is nearly upon us!

If you are a student (or postdoc) working on phylogenetics, taxonomy or systematics, join us Nov 14th for this free and friendly online conference!

Abstract submission closes Saturday, October 25th at 11:59PM GMT+1

REGISTER: tinyurl.com/2w5n3e8
YSF2025 Registration Form
Welcome to the YSF2025 Registration and Abstract Submission form. This year the conference will be held online on Friday, November 14th 2025. We look forward to seeing old and new faces from across t...
tinyurl.com
Reposted by Ana Serra Silva
systassn.bsky.social
Don't forget to register for our (free) upcoming Founder's Lecture on Friday, Oct 24 6PM at Burlington House, London.

Prof @emmateeling.bsky.social, University College Dublin, will be talking to us about all things bats! 🦇

Registration: tinyurl.com/28u2u2z8

Details: systass.org/founders-lec...
Reposted by Ana Serra Silva
systassn.bsky.social
Join us Nov 14th for #YSF2025, a completely free, online conference where postgraduate students (and other early career researchers) present their work on systematics, taxonomy and/or phylogenetics.

Abstract submission closes Oct 25th!

REGISTER: tinyurl.com/2fvmmthc

systass.org/young-system...
Young Systematists’ Forum
Young Systematists’ Forum An exciting setting for postgraduate students and early postdoctoral researchers to present their work to a scientific audience interested in taxonomy and systematic…
systass.org
anaserrasilva.bsky.social
Disclaimer: I am not associated with the journal. I have, however, been press ganged by a gaggle of (friendly but social media-shy) editors to share the articles in this special issue!

Any oversimplifications of the articles' contents are entirely my fault!
anaserrasilva.bsky.social
First up we have Xiyun Jiao and @zihengyang.bsky.social showing the importance of moving beyond model fit testing and embracing sensitivity testing in phylogenetics.

In other words, question your data assumptions and test for model robustness not just adequacy.

(2/n)

doi.org/10.1093/evol...
On measures of influence and discordance in phylogenomic analyses
Abstract. As models are approximate descriptions of real biological processes, and data are collected often with errors and contamination, sensitivity or r
doi.org
anaserrasilva.bsky.social
If you're interested in understanding discordance in phylogenomic analyses, the @evojlinnsoc.bsky.social's special issue 'Phylogenomic Discordance: Patterns, Processes, and Solutions' is for you!

tinyurl.com/v2eces3s

I'll be sharing a few articles a week until we're through the issue! (1/n)🧪
Phylogenomic Discordance: Patterns, Processes, and Solutions
Phylogenomics, the study of evolutionary relationships using genomic data, has revolutionized our understanding of the Tree of Life. As a field, phylogenomics h
academic.oup.com
Reposted by Ana Serra Silva
systassn.bsky.social
We are delighted to announce that this year's Founders' Lecture will be delivered by Prof. @emmateeling.bsky.social, University College Dublin!

TITLE: Bats: Genomes, phylogenies, fossils and cures?

WHEN: Friday, October 24th at 6PM

WHERE: Burlington House, London

JOIN US: tinyurl.com/28u2u2z8
Founders' Lecture details and registration QR code.
anaserrasilva.bsky.social
Always nice to see an old friend. Now to find what shenanigans are afoot this time. 🤓
Reposted by Ana Serra Silva
petermulhair.bsky.social
Last chance to join us next month for the #Biodiversity25 conference.

A fantastic set of speakers, mix of talks & focused workshops, and the chance to meet others in the field of biodiversity genomics 🧬

In-person registration closes next Monday 29th September!

Registration 🔗: bit.ly/4j3vAZ9
Understanding Life: Using Largescale Biodiversity Reference Genomes — 20251027
Understanding life: Using largescale biodiversity reference genomes
bit.ly
Reposted by Ana Serra Silva
systassn.bsky.social
We are nearly ready to announce this year's Founders' Lecture speaker!

In the meantime time check out @maxjtelford.bsky.social's 2024 lecture: 'Are we really more closely related to starfish than to earthworms?'

www.youtube.com/watch?v=FSR5...
Founders' Lecture 2024 | Are we really more closely related to starfish than to earthworms?
YouTube video by Systematics Association
www.youtube.com
Reposted by Ana Serra Silva
maxjtelford.bsky.social
My review of the evidence (morphological, embryological and molecular) for the deuterostomes clade is now available. www.annualreviews.org/content/jour...
Figure showing sagittal sections of a cephalochordate and a hemichordate highlighting potentially homologous structures
Reposted by Ana Serra Silva
anaserrasilva.bsky.social
Not sure I like the look of these clouds...
anaserrasilva.bsky.social
Where are these from? (Assume not NYT, since I don't remember that clue from this morning.)
anaserrasilva.bsky.social
Ah, the one time I want a delivery to be late, it is on track to arrive ahead of time... Guess I'll be starring at the box for the better part of a month 🛥️
anaserrasilva.bsky.social
I'm disappointed it took me so long to notice something rather apt...

So, I abbreviated Bos taurus to BOSTA (phylip format and all), which in my first language just so happens to mean cowpat...

Not that anyone uses its literal meaning, we mostly use it as the mildest faeces-related swear word 😂
anaserrasilva.bsky.social
TLDR: Support for deuterostome monophyly comes from sources of systematic error (branch-length, site-rate and site-compositional heterogeneity). When these are explicitly modelled, we cannot distinguish between monophyletic and paraphyletic Deuterostomia. (12/12)
Phylogenetic tree showing Chordata (fish), Ambulacraria (acorn worm) and Protostomia (butterfly and snail) in a polytomy.
anaserrasilva.bsky.social
AND under the best-fit model, CAT-Poisson (CAT-GTR did not converge), we couldn't distinguish between the Deuterostomia, Orthozoa and Centroneuralia topologies for the short-branched replicate.

Thus, when we mitigate sources of systematic error, support for Deuterostomia nearly disappears. (11/12)
anaserrasilva.bsky.social
Lastly, we ran leave-one-out cross-validations on one short- and one long-branched taxon jackknife under the site-homogeneous LG+G, site-heterogeneous EDM+G and CAT-Poisson+G and found that as model complexity increased, support for deuterostome monophyly.

(10/12)
anaserrasilva.bsky.social
We also found that support for deuterostome monophyly is highest at highly compositionally constrained sites (number of effective AAs < 3).

(We used some simulations to test this further but they don't really fit in 300 chars, so bug me if interested to know more.)

(9/12)
Boxplots showing difference in support between pairs of topologies binned by level of site-compositional constraints (number of effective amino acids). Deuterostomia (purple) always preferred to paraphyletic alternative but much less strongly under a site-heterogeneous model (EDM) and in teh absence of ong-branched taxa. Negligible difference in support between the two paraphyletic alternatives (Orthozoa-salmon and Centroneuralia-green).
anaserrasilva.bsky.social
2) a site-homogeneous model with rate variation

3) a site-heterogeneous model with rate variation

And found that support for deuterostome monophyly is higher when:

1) rate variation is not accounted for

2) long-branched taxa are sampled, and

3) compositional heterogeneity isn't modelled (8/12)
Plot showing that number of sites supporting deuterostome monophyly (purple) decreases when long-branched taxa are excluded and when site-heterogeneous models are used.