Nik Tzoumas

Power is a major confounder in the analysis of cross-ancestry 'portability' in human eQTLs https://www.biorxiv.org/content/10.64898/2026.02.26.708346v1

With some help from Claude Code, I have the app I've always wanted:

elicitcausal lets you design a causal graph with your theoretical priors & preregister it. Then after you complete a study, you can upload your graph and get estimates of causal learning.

Link: causal.wilddata.solutions

#rstats

Cool paper alert. Was trying to model the thermal performance curves of several farmed seaweed species and ran across this awesome, theoretical unifying paper: “A universal thermal performance curve arises in biology and ecology” www.pnas.org/doi/10.1073/...

#rstats Here's a useful guide to creating publication-ready #ggplot figures to journal specifications, which is often quite fiddly.

jaquent.github.io/2026/02/crea...

The lack of uptake of sequential analysis shows how irrational scientists are, and how their methods are driven by norms. Sequential analyses give you more flexibility and are more efficient than a single hypothesis test, and yet, they are still very rare. lakens.github.io/statistical_...

Plot and fit at the same time with a logit dot plot: bsky.app/profile/solo...

Learning lifetime disease liability reveals and removes genetic confounding in electronic health records Electronic health records (EHRs) have become the cornerstone of population-scale genetic studies1, but factors including patterns of healthcare use shape which and how diagnoses are recorded, leading ...

Our new preprint “Learning lifetime disease liability reveals and removes genetic confounding in electronic health records” is now online! Link to paper: This work is led by my postdoc Yazheng Di and it’s our first project at @bsse.ethz.ch :) medrxiv.org/cgi/content/...

Thread 1/n

How far back in time can you understand English? An experiment in language change

If you liked this experiment, I published a full piece today in the same vein: a text that gets 100 years older with every section, from a modern blog post to a medieval chronicle.

It's a single story spanning 1000 years of English. See how far you get.

www.deadlanguagesociety.com/p/how-far-ba...

"The case representing a realistic psychology field study, in which researchers have limited control over variables, stabilized at n = 3,800, requiring 72% statistical power."

#stats #PsycSci

ReverseGWAS identifies combined phenotypes associated with a genotype in GWA studies AbstractMotivation. Traditional genome-wide association studies (GWAS) aim to uncover the genetic variants associated with a single phenotype of interest (

ReverseGWAS identifies combined phenotypes associated with a genotype in GWA studies academic.oup.com/bioinformati... 🧬💻🧪 #Rstats pkg github.com/Leonardini/r...

We hoped this would take 5 years, but in fact it took 14! Thanks to all co-authors, but particularly Mark Jobling who held it all together, and more. A proper magnum opus. Hope it is useful for those in the field or adjacent. Phew.

adrftools: Estimating, Visualizing, and Testing Average Dose-Response Functions Facilitates estimating, visualizing, and testing average dose-response functions (ADRFs) for characterizing the causal effect of a continuous (i.e., non-discrete) treatment or exposure. Includes suppo...

I'm so excited to announce the first release of my newest #Rstats package, {adrftools}! This package facilitates estimation, visualization, and testing for the causal effect of a continuous (i.e., non-discrete) treatment.

🧵 1/10

#statssky #episky #causalinference

Comparative Efficacy and Safety of Ketamine Versus Midazolam for Suicidality: A GRADE‐Assessed Systematic Review and Meta‐Analysis of Randomized Controlled Trials Ketamine rapidly reduced suicidal ideation and depressive symptom severity more than midazolam in adults with acute suicidality across 10 RCTs (n = 649). Transient adverse effects (e.g., nausea, dere...

1. Another problematic systematic review found in my weekly PubMed alert. Ketamine vs. midazolame (=active placebo).
Authors conclude that "ketamine was linked to a considerable decrease in suicidal ideation in comparison to midazolam with mean differences of −1.23 points on the " MADRS

1/2 Nail in the coffin of dichotomania: @erik-van-zwet.bsky.social 's paper with @stephensenn.bsky.social and myself just published: onlinelibrary.wiley.com/doi/10.1002/... with extended discussion at discourse.datamethods.org/t/dichotomiz... #Statistics #StatsSky #rct #clinicaltrial

Wow this is a breakthrough in validation of predictive models that fully accounts for point estimates used in predictions being only point estimates. Great work @richarddriley.bsky.social and colleagues! #Statistics #StatsSky

One approach to the age-period-cohort problem: Just don’t. Just to cause yourself more problems, you seek for something. But there is no need for you to seek anything. You have plenty, and you have just enough problems. Shunryū Suzuki in a 1971 talk A ...

New blog post about the age-period-cohort identification problem!

In which, for the first time ever, I ask "What's the mechanism?" and also suggest that sometimes you may actually *not* be interested in causal inference.

www.the100.ci/2026/02/13/o...

Bluesky Map Interactive map of 3.4 million Bluesky users, visualised by their follower pattern.

I made a map of 3.4 million Bluesky users - see if you can find yourself!

bluesky-map.theo.io

I've seen some similar projects, but IMO this seems to better capture some of the fine-grained detail

Another Claude project: a static site that pulls in GWAS SNP data from ensemble, multiple public biobanks, open targets, gtex, eqtl catalog, and OMIM.

sashagusev.github.io/gwas_lookup/

Mendelian Randomization is powerful but often debated. We benchmarked it against the "ground truth" of human metabolic pathways to see how often it gets it right. 🧬

Result: MR is robust! It matches known metabolism in 43% of cases and identifies valid drug targets.

www.cell.com/ajhg/fulltex...

While we are on this topic:

*when you use logs for data, don't do the natural log! no one cares about units of e!* 🫠

*natural logs are only useful for computation* 🤖

*instead, use a percentile base like log base 1.1. then your covariate is in units of 10%* 😎

*or quadratic: X + X^2* 🫡

#rstats

Sifting through Animal Behavior Society newsletters and came across a castigation of the field made by Dutch ethologist Adriaan Kortlandt in 1990. He states, "Ethology is supposed to be the study of behaviour, but the behaviour has become invisible." He also used to fight goats during lectures.

blog post: 5 flavors of calibration

2 from survey statistics
1 from machine learning
2 from Gelman et al.'s workflow article

this is how I was taught fixed and random effects and I think it still holds up well in practice

Shocking. The argument is so simple but surprising at the same time. I would never have thought about this in practice. Another reason why we should be very careful when operating under "previously established protocols".

Animals experience joy. Scientists want to measure it Scientists have long focused on quantifying fear and other negative emotions in animals. Now they’re trying to measure positive feelings — and it’s a challenge.

A lovely read about researchers who make animals happy for a living 😊 A day spent playing the sounds of baby bonobo laughter is surely a day well spent

www.sciencenews.org/article/scie...

Finally listened to this lecture -- highly recommended if you've ever been confused about fixed effects vs. random effects models. Love that the framing starts from group-level confounding -- here's a problem, what could be solutions (rather than: here's a model, what are its features).

Why are UK universities failing? - LSE Impact The HE sector in the UK faces the prospect of a university going into administration. How have universities fallen so low and is change possible?

This is good on what's it like in a UK uni these days: Why are UK universities failing? - LSE Impact share.google/NYPBoQNH1DLH...

Multiple protein structure alignment at scale with FoldMason Protein structure is conserved beyond sequence, making multiple structural alignment (MSTA) essential for analyzing distantly related proteins. Computational prediction methods have vastly extended ou...

FoldMason is out now in @science.org. It generates accurate multiple structure alignments for thousands of protein structures in seconds. Great work by Cameron L. M. Gilchrist and @milot.bsky.social.
📄 www.science.org/doi/10.1126/...
🌐 search.foldseek.com/foldmason
💾 github.com/steineggerla...

Does the Use of Crowdsourced Listeners Yield Different Speech Intelligibility Results Than In-Person Listeners for Typically Developing Children? Purpose: We examined the performance of crowdsourced listeners compared with in-person listeners on the measurement of speech intelligibility for...

if you want to know how to compute marginal means from a mixed effects model in brms using rvars, then have I got a supplemental materials document for you pubs.asha.org/doi/10.1044/...

Why evidence matters in health research Professor George Davey Smith explains why he encourages scientists to look for ways they might be wrong.

Mistaking correlation for causation doesn’t just waste research funding, it can shape real-world health advice.

In our new blog, Prof. George Davey Smith, 2025 MRC Millennium Medal winner, explains why he encourages scientists to look for ways they might be wrong.

www.ukri.org/blog/why-evi...