Gautam Shirsekar

18 Followers
50 Following
14 Posts
I am a plant pathologist. I study host-pathogen coevolution in wild plant pathosystems using population genomics. (Assistant Professor, Entomology and Plant Pathology, at Univ. of Tennessee, Knoxville) www.coevolutionlab.org
Wild Flowers Festival at Frozen Head State Park #stateparks
#TN with Tennessee Native Plant Society (Larry Pounds)
#TNPS
Frozen Head State Park, Tennessee
Frozen Head State Park, Tennessee
On Friday,
@diazale successfully defended a wonderful thesis on modelling high-dimensional genetic data. A must-read if you are working with UMAP or genetic clustering, or if you just like well-written and thoughtful scientific writing. Congratulations!
@alxsim @coevolution
You can find it here: https://github.com/diazale/phd_thesis.
My understanding is that the S t-SNE refers to the fact that we are considering distances between neighbouring points as drawn from a distribution, and we are trying to align low- and high-dimensional distributions. In principle one could use this to estimate uncertainty in representation, but the truth is that the "best" representation is so hard to find, due to many local minima, that other sources of uncertainty dominate.
GitHub - diazale/phd_thesis: PhD thesis for Alex Diaz-Papkovich

PhD thesis for Alex Diaz-Papkovich. Contribute to diazale/phd_thesis development by creating an account on GitHub.

GitHub
Use of t-SNE and #UMAP is increasing in biology! So I wonder how stochasticity of both the algorithms affect how we interpret the results? Can somebody enlighten? (#tSNE at least makes it clear in its acronym that it is stochastic)

I'm enjoying this new preprint by Bénitière, et al on alternative splicing (AS) and the drift barrier

Panel A: abundant AS forms tend to be frame-preserving (i.e., sensible), whereas the rare ones look like a lot of junk

Panel C and D: the two kinds have opposing correlations with proxies for 1/N_e (longevity and body size)

https://www.biorxiv.org/content/10.1101/2022.12.09.519597v5

Yep.

Comic by Sarah Andersen - https://sarahcandersen.com/

#books

Sarah's Scribbles

Tumblr

⚠️ Science / evolution chums ⚠️

If you want a spatially explicit eco-evoutionary simulation to play with, we have just released a new version of REvoSim:

https://joss.theoj.org/papers/10.21105/joss.05284

It runs (with GUI) on all operating systems. Binaries here:

https://github.com/palaeoware/revosim/releases

#science #evolution #nature #software #ecology

REvoSim v3: A fast evolutionary simulation tool with ecological processes

Furness et al., (2023). REvoSim v3: A fast evolutionary simulation tool with ecological processes. Journal of Open Source Software, 8(89), 5284, https://doi.org/10.21105/joss.05284

Journal of Open Source Software
Big Biology - Long-term experimental evolution in the wild (Ep 106)

Can we predict evolutionary outcomes if we know starting conditions? Do the products of evolution in nature differ from those studied in well-controlled lab experiments? On this episode, we talk to Katie Peichel, head of the Division of Evolutionary Ecology at the University of Bern, Switzerland, and Andrew Hendry, professor in the Department of Biology at McGill University, Canada. Katie and Andrew are part of a massive research team working on the evolution of threespine sticklebacks as they are reintroduced into lakes in Alaska. Sticklebacks have been a favorite species for evolutionary biologists since almost the origins of modern evolutionary theory. Traits like spine size and lateral plate armor evolve rapidly when populations colonize new habitats, leading populations to barely resemble one another. Unlike traditional evolutionary experiments, which try to infer what occurred in the past, the Alaska project is tracking in unparalleled detail changes in the phenotypes and genotypes of fish that went into each lake population. We talk to Katie and Andrew about the origins of this incredible project, the pros and cons of different approaches to studying evolution, and the need for long-term experimental studies of evolution in the wild. This is the first of a series of episodes we will be doing on the Alaskan research project, so stay tuned! Cover art: Keating Shahmehri

Google Podcasts