Despite this being true, I feel that a lot of gems from the old days are lost in the present literature. People just do things as if those early findings don't exist.

Within-family heritability estimates for behavioural and disease phenotypes from 500,000 sibling pairs of diverse ancestries https://www.medrxiv.org/content/10.1101/2025.09.17.25336022v1

SimHumanity: Using SLiM 5.0 to run whole-genome simulations of human evolution https://www.biorxiv.org/content/10.1101/2025.09.01.673541v1

PCA for visualising large #pedigrees? Check out this brief and accessible blog post wp.me/pbTxyW-1ve @roslininstitute.bsky.social @highlanderlab.bsky.social Thanks to co-authors @epigenci.bsky.social @gregorgorjanc.bsky.social and Ros Craddock!

Save the date! #PEQG26 June 9-12 2026 in Asilomar, CA. Happens only every 2yrs, but is my favorite conference. Full website coming soon, and registration and abstract submission opens November 14, but I'm allowed to tease that keynotes will be @jnovembre.bsky.social @jennytung.bsky.social and me!

Announcing SimHumanity, a baseline SLiM 5.0 model of the full human genome, replete with demographic history, autosomes, X/Y, and mtDNA. A shared starting point for reproducible evolutionary simulations. We’d love your feedback! #SLiM #evolution #genomics www.biorxiv.org/content/10.1...

Our pedigree PCA article is now online on GSE: doi.org/10.1186/s127... Thanks @epigenci.bsky.social , @gregorgorjanc.bsky.social , Rosalind Craddock!

This always blows my mind. Dude mapped qtl for a quantitative trait to a chromosome in 1923!

I'll be advertising a couple of PhD scholarships (breeding/wood quality, breeding/remote sensing) in the next couple of weeks, working with me and even cooler people at the School of Forestry, University of Canterbury, Christchurch, New Zealand. They cover uni fees + stipend for 3 years.

Assessment of segregation variance estimates from derivation, simulations, and empirical data in autotetraploid species exemplified in potato https://www.biorxiv.org/content/10.1101/2025.08.13.670171v1

AlphaSimR course edin.ac/3wfGSEj Week 3 opens with a discussion on genetic variation between relatives, with a focus on siblings. While mutation is the ultimate source of genetic variation (by introducing new alleles), recombination shuffles the existing genetic variation through meiosis.

Deep-learning-based gene perturbation effect prediction does not yet outperform simple linear baselines www.nature.com/articles/s41... 🧬🖥️🧪 #Rstats code github.com/const-ae/lin...

You can save #tidyplots in multiple formats by piping through multiple rounds of `save_plot()` 🤩 #rstats #dataviz #phd

Looks like a great resource for teaching popgen inference.

Thanks! This is amazing! So the term 'Manhattan Plot' is not originally a GWAS term at all. Screenshotting the image here from a 1994 book on nuclear physics, for others who may be interested:

This is how you can visualize time course data in #tidyplots ⏱️ #rstats #dataviz #phd

Does anyone know who first used the term "Manhattan Plot"? Wikipedia links to a 2010 paper by Greg Gibson. I haven't yet found an earlier use, but Greg's phrasing suggests that he did not invent the term. doi.org/10.1038/ng07...

Can anyone point me to full resolution versions of these classic images from NHGRI that show the growth of GWAS from ~2006--2015? I'm looking for the timelapse versions. (I know many of you are waiting for the GWAS chapters of my free online textbook on human genetics--they are coming, but slowly!)

Very pleased to say we have an Assistant Professor job opening (tenure track) in the Department of Genetics at Trinity College Dublin! 😀 www.jobs.ac.uk/job/DOA399/a... - deadline is August 19th! Please repost!

Our models show that continued imports could alter the genetic makeup of native populations threatening local adaptations. We propose a proactive solution: Invest in selective breeding of the native Irish honey bee to improve key traits within a population already adapted to local conditions. (5/6)

🟢Non-native genes spread rapidly, reaching colonies over 80 km away from the import location. 🟢Halting imports did not fully reverse the trend —introgression persisted for generations. (4/6)

Our key findings: 🔍 🟢Even small percentages of imports (1.5–4%) resulted in notable introgression of non-native alleles. 🟢While honey yield improved, there was a consistent decline in genetic fitness under Irish environmental conditions. (3/6)

We used stochastic simulations to study how non-native genes spread over time and affect fitness and honey production. Find it here: 👇👇👇 doi.org/10.1101/2025... (2/6)

🐝New Preprint!🐝 “Modelling the impacts of imports of non-native honey bees into the native Apis mellifera mellifera population in Ireland”. (1/6)

Hence, this highlights the fundamental difference btw genomic and family h2. The former is considering millions of years of mutation while the latter is looking at everything but mutations over a very short period of time.

Now, if genomic mixed models are implicitly taking the underlying ARG into account, it's considering mutations that accumulated over millions of years. At the time, various disequilibria like LD are absent because the effect of recombination was conditioned out due to conditioning on the ARG.

Then why on earth is the branch effect random? That's because mutations hit branches randomly. Everytime mutations realize, branch exhibits different effect sizes. The consequence is that we get everything about the random effects, like independence, for free without any further assumptions.

My paper argues that what people have considered as SNP random effects b~p(b) is in fact the effect of a *branch* of the ARG in which the SNP belongs to. All SNPs on the same branch cannot be distinguished anyway. Hence, only their aggregate effect (=branch effect) is identifiable.

This is from my personal website. This time, the prior was driven by mutation. More examples coming soon.

Surprisingly, branch genetic relatedness derived from inferred trees are usually off only by a constant from the true value. Hence, the mixed model with inferred trees does just as good as the true ARG, as shown demonstrated by a spatial simulation.

The runtime is linear w.r.t. sample size. How? we used the new matvec algorithm presented in an earlier preprint by Brieuc Lehmann. I designed a second order Newton update that is almost as cheap as a first order gradient descent. www.biorxiv.org/content/10.1...

Pls read the paper for more detailed arguments. Now we present `tslmm`, a linear REML algorithm built on top of the tskit ecosystem. github.com/hanbin973/ts...

(3) Where does genetic confounding come from? We show that fixed effects covariates don't exist when the evolution is neutral: YOU DONT NEED TO ADJUST FOR PCs if you think the evolution was indeed neutral (although we know it's not true in practice).

(2) Pseudoreplication is inevitable in estimating genetic signals. Even if the two cohorts are disjoint, they still have a shared ancestry that leads to dependent parameter estimates. The following is an example of variance component estimates from two non overlapping samples.

So what do we learn from the theoretical argument? (1) The underlying demography governs the upper bound of polygenic prediction accuracy. Why? more inbred/bottlenecked populations have higher genetic variance.

ARG-LMM isn't merely a statistical model that estimates numbers. It's about how traits actually emerge from an evolutionary process. Hence, we obtain deep insights about what we can actually learn from data.

This generative process we call it the Ancestral Recombination Graph Linear Mixed Model, ARG-LMM. Everything about the model is specified as a function of the underlying ARG without arbitrary statistical assumptions.

We show that random effects are driven by random mutations landing on a fixed genealogy. This immediately proves that (1) random effects are independent (2) spike-and-slab prior is the most sensible (3) the weights are a function of ARG's shape (branch lengths, span etc.)

My talk from this year's probgen is finally out. Where do linear mixed models and random effects come from? They emerge from *mutations* on Ancestral Recombination Graphs.

Richard Sever explaining costs of bioXiv - it's so cheap! 10s of $ per article, total annual costs less than one typical genetics lab. this is two orders of magnitude less than costs of journal per article.

New preprint: "Genomic models for accurate estimation of breeding values in Sitka spruce (Picea sitchensis Bong. Carr)" from the Sitka Spruced project www.researchsquare.com/article/rs-7...

1/ Can you pay to get taller, smarter embryos? Some companies are already offering parents pursuing IVF the chance to "see and understand the complete genetic profile of each of their embryos.” But does embryo selection work? Not so much... open.substack.com/pub/jenndowd...

Our new work on estimation of background selection (BGS) is up on bioRxiv. We investigated the problem using simulations and also developed a method for joint inference of BGS, demography and partial selfing from windowed site frequency spectra along the genome. Feedback is welcome. #BGS

Sad and exciting day for large-scale genomics research:(

Two 3y postdocs available at Edinburgh's Roslin Institute in breeding *simulation software development* and *biometrics/applied statistics*. Check out elxw.fa.em3.oraclecloud.com/hcmUI/Candid... and elxw.fa.em3.oraclecloud.com/hcmUI/Candid... @gregorgorjanc.bsky.social @roslininstitute.bsky.social

Here is one big reason to choose Europe to advance with your research career: We 💙 science.

Elegant new theory on polygenic adaptation by Will Milligan, Laura Heyward, and Guy Sella.

How to estimate the additive genetic variance for fitness, without measuring fitness. Work from Manas Geeta Arun and Jarrod Hadfield: www.biorxiv.org/content/10.1...

Can recommend both the conference and the city 👇