eQIF predictions for common European genomic variants

This dataset provides genome-wide precomputed eQIF scores for 9,421,183 common human single-nucleotide variants (SNVs) in GRCh38/hg38 coordinates. eQIF uses an unsupervised isolation forest trained on the top 20 Enformer principal components to prioritize variants with potential gene-regulatory effects. Columns: chr: Chromosome pos: Genomic position (hg38) ref: Reference allele alt: Alternative allele eQIF: Raw isolation forest anomaly score eQIF_quant: Genome-wide quantile of the eQIF score Higher eQIF scores indicate greater predicted gene-regulatory potential. The eQIF scores were derived from the publicly available Enformer variant-effect predictions (top 20 principal components) provided by Avsec et al. (2021), licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). These predictions were integrated with gnomAD variants, converted to GRCh38/hg38 coordinates, and processed using an isolation forest to generate the eQIF scores. Original Enformer data: https://console.cloud.google.com/storage/browser/dm-enformer/variant-scores/1000-genomes/enformer/PC20