Lightnews — Scholar-powered news

Peter Kraft

@peter-kraft.bsky.social

32 followers 28 following 30 posts

Cancer epidemiologist, statistical geneticist, biostatistician. National Cancer Institute, Harvard. Views my own.

Posts Media Videos Starter Packs

Peter Kraft @peter-kraft.bsky.social · 11h

Speaking of ripple effects: any guidance for NIH researchers who have registered for #ASHG25 and have a poster or presentation but may not be able to attend because of the shutdown?

Reposted by Peter Kraft

American Society of Human Genetics (ASHG) @geneticssociety.bsky.social · 2d

@ajhgnews.bsky.social sat with Julie-Alexia Dias, MSc, in the latest "Inside AJHG" to discuss her recently published paper, “Evaluating multi-ancestry genome-wide association methods: statistical power, population structure, and practical implications.”➡️ ashg.org/ajhg/inside-... #ASHG #humangenetics

Julie-Alexia Dias, MSc

Peter Kraft @peter-kraft.bsky.social · 14d

I see this playing out in slow motion, accompanied by Barber’s Adagio.

Peter Kraft @peter-kraft.bsky.social · Sep 8

Curious to hear others’ thoughts and experience here! /fin

Peter Kraft @peter-kraft.bsky.social · Sep 8

(iv) And the simulations and applications assume individual-level data or in-sample LD is available—typically not the case in large meta-analyses for complex traits. See Wenmin Zhang et al for a discussion of this issue and a possible fix. www.biorxiv.org/content/10.1... 12/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

On a prosaic level, defining discrete genetic ancestry clusters often means excluding participants who don’t fall into any of the clusters, lowering sample size and limiting generalizability. 11/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

(iii) As the authors note, discrete genetic ancestry groups are made-up things. There are a bazillion ways to define clusters of participants by projecting their genotypes into some abstract mathematical space—it’s not clear which (if any) adequately captures variation in genetic effect. 10/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

(ii) The simulations are focused on two ancestry groups, with imbalance maxing out at 1:2. In the complex trait setting, it’s not unusual for there to be 4 or 5 ancestry groups, with imbalances on the order of 1:5 or more. 9/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

Some caveats and open Qs: (i) The simulations and data applications are focused on the context of molecular QTL (large effects, small sample sizes) not complex traits (small effects, large sample size). Not clear (but plausible) that qualitative results transfer to that setting. 8/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

Points (ii) and (iii) may be the flip side of this: in low power situations, the extra degrees of freedom allowing for group-specific effects may cost power. Betting on near similar effects borrows information across groups and improves power here. 7/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

…then SuShiE and other methods that allow effects to differ across ancestry groups should be more powerful. On the other hand, if genetic effects are nearly identical (e.g. the causal variant is typed and marginalized GxE and GxG effects are negligible) then pooling should be more powerful. 6/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

Point (i) makes sense: if the genetic effects differ across ancestry groups—perhaps due to linkage disequilibrium differences if the causal variant is not typed or due to subtle differences in marginal genetic effects due to GxE and GxG interactions… 5/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

Quick take-homes: SuShiE outperforms pooled SuSiE—except when (i) the correlation in genetic effects across ancestries is very high (0.99), (ii) sample sizes across ancestries are imbalanced, or (iii) the overall sample size is low relative to the strength of the genetic effects. 4/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

We just compared pooled versus stratified analysis of GWAS for locus discovery (pubmed.ncbi.nlm.nih.gov/40902600/), so I was particularly interested in the comparisons of SuShiE and other methods that rely on genetic-ancestry-group-stratified analyses to SuSiE applied to the pooled data. 3/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

If fine-mapping, mol-QTL, or [fill-in-the-blank]WAS analyses are your jam, do check this paper out, if only for the nice review and assessment of contemporary multi-ancestry fine-mapping methods. If fine-mapping is not your jam, this is gonna get technical & jargony. 2/n

Peter Kraft @peter-kraft.bsky.social · Sep 8

This was neat work by @nmancuso.bsky.social et al developing and benchmarking a new multi-ancestry fine-mapping method (“SuShiE”). I learned something about the performance of pooled v stratified analyses but still have some Qs. 1/n

Nature Genetics @natgenet.nature.com · Jul 21

📢OUT TODAY @natgenet.nature.com

📰Improved multiancestry fine-mapping identifies cis-regulatory variants underlying molecular traits and disease risk.

By @zeyunlu.bsky.social, @nmancuso.bsky.social and colleagues.

⬇️

www.nature.com/articles/s41...

Improved multiancestry fine-mapping identifies cis-regulatory variants underlying molecular traits and disease risk - Nature Genetics

SuShiE is a multiancestry fine-mapping method for molecular quantitative trait loci that leverages linkage disequilibrium heterogeneity to improve resolution, infer cross-ancestry effect size correlat...

Peter Kraft @peter-kraft.bsky.social · Sep 3

In epi & biostats it’s not unusual for chapters to be published papers with full author list (often including advisor). If chapter is draft of a paper yet to be shared with coauthors, no author list included (contributions in acknowledgements). Student meets first author criteria in both scenarios.

Reposted by Peter Kraft

James 🎃cello 🌉 @jamespirruccello.com · Sep 2

The observation that joint analysis can safely improve power strikes me as being a similar observation to that of @genandgenes.bsky.social et al in their Nature 2019 manuscript: www.nature.com/articles/s41...

Genetic analyses of diverse populations improves discovery for complex traits - Nature

Genetic analyses of ancestrally diverse populations show evidence of heterogeneity across ancestries and provide insights into clinical implications, highlighting the importance of including ancestral...

Peter Kraft @peter-kraft.bsky.social · Sep 2

Thanks to all coauthors (including @madduri on here) for important substantive and technical contributions. 7/7

Peter Kraft @peter-kraft.bsky.social · Sep 2

A 2019 review by Peterson et al. flagged “to pool or not to pool” as one of the key unresolved issues in the analysis of GWAS in ancestrally diverse samples. We hope our paper provides some guidance. 6/7 www.sciencedirect.com/science/arti...

Genome-wide Association Studies in Ancestrally Diverse Populations: Opportunities, Methods, Pitfalls, and Recommendations

Genome-wide association studies (GWASs) have focused primarily on populations of European descent, but it is essential that diverse populations become…

www.sciencedirect.com

Peter Kraft @peter-kraft.bsky.social · Sep 2

These are encouraging results, given the ongoing expansion of GWAS to include more diverse samples (e.g. the Confluence Project, which will double the sample size of the largest breast-cancer GWAS to date and more than quadruple the number of Latina participants). 5/7 confluence.cancer.gov

DCEG Data Platform

confluence.cancer.gov

Peter Kraft @peter-kraft.bsky.social · Sep 2

Also: trust the maths, but verify. A pooled analysis is a good place to start, but do your due diligence and check lambda-GCs and LDSC intercepts for evidence of concerning inflation. Assessing the correlations between the trait of interest and global and local PCs can also help. 4/7

Peter Kraft @peter-kraft.bsky.social · Sep 2

Your mileage may vary: power gains and confounding are trait- and context-specific. Our simulations and real-data examples are focused on anthropometry, biomarkers, and complex diseases. Confounding could be more of an issue for socially defined traits (e.g. educational attainment). 3/7

Peter Kraft @peter-kraft.bsky.social · Sep 2

Through a combination of maths, simulations, and real-data analyses, we show that pooled analysis is generally more powerful than meta-analysis while controlling Type I error rates. 2/7

Peter Kraft @peter-kraft.bsky.social · Sep 2

Multi-ancestry GWAS can increase power and precision, but how should we analyze them? Pooled or stratified? We answer that question in a paper out today in AJHG, led by Julie Dias and Haoyu Zhang. 1/7 www.cell.com/ajhg/fulltex...

Evaluating multi-ancestry genome-wide association methods: Statistical power, population structure, and practical implications

Multi-ancestry GWASs enhance discovery in diverse populations, but optimal methods remain debated. Using theory, simulations, and analyses from the UK Biobank and All of Us, we show that pooled analys...