Functional annotations integrated into CAUSALdb can be divided into four major categories according to the attributes and usage of the collected datasets: variant information, functional prediction, functional evidence, and trait association. Variant information annotations report basic information, including variant genomic features derived from dbSNP and CADD (31) and variant allele frequency derived from gnomAD (32) and 1KGP phase 3 (33). Functional prediction annotations incorporate several integrative variant function prediction results from frequently used resources, including aggregated conservation scores from CADD, aggregated non-coding variant prediction scores from regBase (34), aggregated missense mutation pathogenic scores from dbNSFP (35), aggregated splicing altering prediction scores from dbscSNV (36), aggregated miRNA-target altering prediction scores from dbMTS, and several function predictions from HaploReg (37), RegulomeDB (38) and InterVar (39). Functional evidence annotations integrate large-scale tissue/cell type-specific epigenomic profiling (e.g. histone modifications, transcription factor binding, open chromatin and nascent transcription) data from different resources, such as the Roadmap Epigenomics Project (40), CistromeDB (41) and FANTOM5 Project (42). Trait association annotations collect important disease/trait-associated information, including GTEx eQTLs (43), GWAS Catalog significant variants, ClinVar reported variants (44), DisGeNET recorded variants (45), and ICGC somatic mutation information (46) (Supplementary Table S3).
