aFCn for independent cis-eQTLs across 49 tissues in the GTEx v8 release

The aFC-n model is the multi-variant generalization of the aFC approach (https://github.com/secastel/aFC), to estimate the cis-regulatory effect size in genes associated with multiple conditionally independent eQTLs. Notably, aFC-n is the first method that allows for predicting genetically regulated gene expression in a haplotype-specific fashion, paving the way for a future class of genome association studies that can systematically incorporate allele dosage effects.

This dataset is generated using aFC-n software package (10.5281/zenodo.8412460) to calculate effect sizes for ~half a million independent cis-eQTLs across 49 tissues in the GTEx v8 release as described in the manuscript.

GTEx_v8_aFCn contains:

• gene_id : Ensembl gene ID
• variant_id : eQTL ID associated to the gene
• log2_aFC : The log2 of eQTL effect sizes measured as allelic Fold Change (aFC).
• log2_aFC_min_95_interv : Lower bound 95% confidence interval
• log2_aFC_plus_95_interv: Upper bound 95% confidence interval

GTEx_v8_aFCn_combined contains:

• gene_id : Ensembl gene ID
• variant_id : eQTL ID associated to the gene
• rest of the columns : The log2 of eQTL effect sizes measured as allelic Fold Change (aFC) in the tissue specified by the column label capped at ±log2(100) 
  • nan = The variant is not an eQTL for the tissue / the effect size is not calculated for variants on chrX