Mastodawn

Marc Robinson-Rechavi Feb 14, 2025

Excellent talk by Isabela do O #Biology25, PhD student with @jgx65, showing how population structure can bias Qst-Fst selection tests, and how her new LogAV method can correct and account for population statistics. #PopulationGenetics #MethodsMatter

Marc Robinson-Rechavi Feb 7, 2025

Excited to read this, I've been worried about the recurrence of these peaks in #PSMC graphs for years. Also, just on time before we submit one (a first for me). #PopulationGenetics #PopulationGenomics #MethodsMatter
https://bsky.app/profile/leonhilgers.bsky.social/post/3lhl5ti3nec25

(If Mastodon doesn't show the Bluesky post: "Avoidable false PSMC population size peaks occur across numerous studies" https://doi.org/10.1016/j.cub.2024.09.028)

Leon Hilgers (@leonhilgers.bsky.social)

‼️These #PSMC #peaks are #artifacts‼️ Seen #PSMC results like this🤔? They are caused by ill-set parameters, #very #common, and #easy #to #solve! For details see 🧵& OA paper @currentbiology.bsky.social 🙂🐢🧬🎉! https://cell.com/current-biology/fulltext/S0960-9822(24)01239-9

Bluesky Social

Marc Robinson-Rechavi Aug 16, 2024

A colleague has attracted my attention to the recent Snyder lab paper on two major periods of aging in humans and to Lior Pachter’s reaction on TwiX. Apart from the risk of over-fitting noisy data, is it just me or do they use “linear/nonlinear” to mean “monotonous/non-monotonous”?
https://www.nature.com/articles/s43587-024-00692-2
https://threadreaderapp.com/thread/1824338306391482767.html #MethodsMatter #Aging #Omics

Nonlinear dynamics of multi-omics profiles during human aging - Nature Aging

Understanding the molecular changes underlying aging is important for developing biomarkers and healthy aging interventions. In this study, the authors used comprehensive multi-omics data to reveal nonlinear molecular profiles across chronological ages, highlighting two substantial variations observed around ages 40 and 60, which are linked to increased disease risks.

Nature

Marc Robinson-Rechavi Aug 9, 2024

Interesting result, and I'm surprised no one showed it earlier (apparently?): the apparent increase in evolutionary rates over short time scales can be entirely explained by noise, with randomised data providing identical results to the original datasets 😮
https://www.biorxiv.org/content/10.1101/2024.02.29.582777v4 #phylogenetics #EvolutionaryBiology #MethodsMatter

Show thread

Marc Robinson-Rechavi Jan 16, 2024

Previous articles by @caseywdunn and us have shown that phylogenetic methods are important to testing the ortholog conjecture but that #MethodsMatter and results can be biased in several ways. We follow on those previous studies by using gene expression as a proxy for gene function, both tissue-specificity and mean level of expression.