Thread by @tuuliel

In addition to my main GTEx thread, I’d like to highlight some of the work & companion papers from my lab by these awesome folks: @skimhellmuth @silvakasela Margot Brandt, @EinsonJ, @EliseScience @stephanecastel.

https://twitter.com/tuuliel/status/1304121094119526400?s=20

The cell type work in GTEx was led by @skimhellmuth who, together with @francoisaguet, computationally estimated the enrichment of 64 cell types in all GTEx samples & mapped cell type interacting eQTLs and sQTLs for 7 cell types. science.sciencemag.org/lookup/doi/10.… science.sciencemag.org/content/369/65…

Cell type composition is really a key factor for expression variation across tissues. We show that cell type interaction QTLs resolve multiple eQTLs for the same gene, and the improved resolution adds many new GWAS colocalizations.

@silvakasela mapped population-biased eQTLs for the main paper, finding a handful where effect size differs between Eur/Afr Americans. She also showed the convergence of disease signal for eQTLs & sQTLs and eQTLs & rare coding variants for the same gene.

Different variants with a different molecular effect on a gene showing a concordant GWAS signal is really exciting: such genes are gold-standard causal genes with multiple lines of info, and indicate potential for allelic series of the relationship b\w molecular vs GWAS effect.

@EliseScience described in the main paper how a breast cancer eQTL for CBX8 seems to be due to a SNP that disrupts EGR1 binding. She also showed that eQTL effect size & eGene expression often (anti)correlate: the strongest eQTL can be in the lowest expr tissue or the opposite.

In a @GenomeMedicine companion, Margot Brandt led the work to experimentally test GTEx eQTLs and nonsense variants. With an allelic polyclonal CRISPR assay, we were able to see eQTL effects and esp nonsense-mediated decay, including in Mendelian genes. genomemedicine.biomedcentral.com/articles/10.11…

@stephanecastel led the tissue-specificity work of the main paper. His thread has more details; one highlight is that sQTLs are more shared than eQTLs - likely because of higher tissue sharing of post-transcriptional effects gene regulation.

https://twitter.com/stephanecastel/status/1304121250399150080?s=20

In a @GenomeBiology companion, @stephanecastel describes the GTEx allele-specific expression data - now with open-access haplotypic counts from phASER. We also describe tools & methods for easier eQTL replication with ASE data, which we used extensively. genomebiology.biomedcentral.com/articles/10.11…

We also contributed to the sex specificity paper, led by Meritxell Oliva and @be_stranger, with @skimhellmuth leading the sex-interacting eQTL mapping. One of the highlights for me was seeing how cell type composition underlies sex interactions. science.sciencemag.org/lookup/doi/10.…

In the paper led by @ferraronm @BennyStrobes @sbmontgom @alexisjbattle and lab alumni @Pejminister, @EinsonJ showed how allelic expression (with expression & splicing outliers) captures rare functional variation. Also validation with our new CRISPR assay! science.sciencemag.org/lookup/doi/10.…

I also just learned that the @GenomeBiology and @GenomeMedicine papers aren't online yet, so you may need to wait a while (they should be up in 24h if not before).