You're invited to register for and send in your submissions to #WCPG26!
The International Society of Psychiatric Genetics' annual congress will be held in Glasgow, Scotland from 29 September - 3 October. The theme - Understanding Today, Translating Tomorrow - asks us to look forward. Join us.
Out earlier this week in @natgenet.nature.com: GWAS of major anxiety disorders in 122,341 European-ancestry cases identifying 58 loci (www.nature.com/articles/s41...)
Awesome work, and great news for the anx genetics community — each @pgcgenetics.bsky.social paper seeds 100s more papers!
www.nature.com/articles/s41... 🧪. Impressive GWAS meta-analysis from the Anxiety Disorders Working Group of the Psychiatric Genomics Consortium
Interested in how mental health develops?
If your background is in psychology, biosciences, or a related field, our MSc in Developmental Psychology & Psychopathology at @kingsioppn.bsky.social could be your next step: www.kcl.ac.uk/study/postgr...
Join us in September – reach out to learn more!
Are you seeking a PhD in genomics or the application of AI to healthcare and therapeutics?
Check out KCL's new PhD programme in Genome Data Science for Therapeutic Target Discovery!
Apply by Saturday 28th February 2026, 23:59 GMT
Details and projects: www.kcl.ac.uk/research/tar...
Latent factors are orthogonal, so fine-mapping is more straightforward because there is no trait trait correlation.
Application to blood cell traits allows identification of relevant biology and causal variants
Latent factor GWAS - flashFMZero. Common underlying biological mechanisms.
pubmed.ncbi.nlm.nih.gov/40220762/
See improved results on lipid analysis using MGflashFM compared to original analysis by relaxing previous assumptions and limitations.
MGflashFM, Multi-group multi-trait fine-mapping. Flexible framework , simple interpretation: www.nature.com/articles/s41...
Multi ancestry methods also allow for improved fine mapping (again, assuming shared causal variants). Older methods exclude non shared variants, but this results in a lot of loss with more ancestries and more diverse ancestries.
Also see resolution improvement on glycaemic traits, even more so when combined with function annotations, as well as higher posterior probabilities for top variants: pubmed.ncbi.nlm.nih.gov/41251494/
Applied to lipid traits in the Ugandan Genome Resource. Can resolve a locus for HDL and LDL in APOE in 6407 participants. See different variant in flashFM versus single trait methods for HDL. In a bigger AFR meta-analysis, the flashFM variant is then supported by single trait fine-mapping.
FlashFM - allows multiple causal variants per trait accounting for correlations. Flexible framework, shared information (without bias — if no shared effects, results are as single trait mapping), simple interpretation.
www.nature.com/articles/s41...
Joint analysis is computationally burdensome - intersections of phenotypes rapidly increase the search space. Need to account for trait selection and define sensible priors.
Fine-mapping is the process of trying to find causal variants from associated ones, a specific example of identifying true effects among correlated variables
Why joint mapping? Uses the proposed shared effects to gain power without needing new recruitment, and allows pleiotropy to be studied.
Finally, Jennifer Asimit will discuss leveraging phenotypic similarity and ancestry diversity in cardiometabolic genetics.
Multi trait and multi ancestry methods for fine-mapping are emerging but challenging to implement
Also coverage issues - short read sequencing leaves gaps, which if not accounted for can be misunderstood. [Maybe long read from UK Biobank will help with this?]
UK Biobank allow you to explore mechanisms - can show proteomic and other omic effects for associations seen at organismal level.
But caveat emptor - interesting results can turn out to be explainable by longer-distant haplotype effects (especially indexed by D' and not by r²) of known effects
Porting methods into multiple biobanks. Adjust each strata of the meta-analysis for the overall top variant, then re-meta-analyse. Result is robust to haplotype effects.
Very important for multiple ancestry analysis. Most effects are suppressed. Still ID interesting results.
Heritability probably the wrong metric for rare variants - not impactful on population level, but probably very important in carriers
Very hard to estimate heritability to very rare variants - confounding via population stratification causes inflation.
Saturate heritability at about MAC>10 (ignoring pop strat)
Substantial aggregation of UTR variants in height for FGF5[?]
Yengo et al saturated the common variant space in height - does rare variation lie near to common variants? Yes, mostly. Independent, colocated signals. For height, rare heritability is located near these loci. Not true for BMI,WHRadjBMI
E.g. IGF2BP2 variants - rare regulatory variants upstream of the gene associated with WHRadjBMI. Gene has been previously associated with this. But profile with other phenotypes is different - regulatory variants don't need to do the same thing as nearby regulatory variants!
Proximal Vs intergenic Vs sliding window annotations of regulatory effects. Integrated with further annotations and conditional on nearby coding variants.
GH developed a framework for single variant association analysis for MAC>5 and genomic aggregate testing. Latter is more straightforward for coding - most affect alleles are going to be deleterious. Non-coding is harder - less of a prior for same effects. Examined different aggregation categories .
Final session - Gareth Hawkes talks about using WGS to examine the convergence of rare and common traits.
Preprint: www.biorxiv.org/content/10.1...
Glad to see this out!
Finally Georgios Kalantzis on a recessive meta-analysis across six biobanks, including diverse samples and multiple phenotypes.
58 sig associations. 17 better fit recessive than additive models. See HBB associations that are partly (but not fully) explained by anaemias.
Examined PIEZO1 variant in Genes and Health GWAS. MAF 3.9% in CSA, negligible elsewhere.
Clinical implications of variation - increased delayed diagnosis, complications.
