Table 1.
Methods used for the gene prioritization, and number of genes they prioritize
| Method | Description | Access | Reference | # of genes prioritized in the 536 BMI-associated loci | # of established obesity genes identified (max 6)a | % of established genes identified relative to total genes prioritized | Standardized weights |
|---|---|---|---|---|---|---|---|
| Locus-based methods | |||||||
| Mutation significance cutoff (MSC) | provides gene-level and gene-specific phenotypic impact cutoff values, as opposed to a single significance cutoff value across all genes | https://lab.rockefeller.edu/casanova/MSC | Itan et al.36 | 235 | 3 | 1.30% | 8 |
| Nearest gene | gene nearest to lead variant | – | – | 547 | 3 | 0.50% | 3.7 |
| Transcriptome-wide association study (TWAS) + COLOC | TWAS leverages expression imputation (pre-computed gene expression weights generated from individuals for whom both gene expression and genetic variation have been measured) to test for significant genetic correlation between cis expression and GWAS; the imputed expression can be viewed as a linear model of genotypes with weights based on the correlation between SNPs and gene expression in the training data while accounting for LD among SNPs; COLOC further estimates the posterior probability of colocalization, where colocalization is defined as one (or more) shared causal variants between the expression and GWAS | http://gusevlab.org/projects/fusion/ | Gusev et al.24 & Giambartolomei et al.22 | 160 | 1 | 0.60% | 3.2 |
| Multi-marker analysis of GenoMic annotation (MAGMA) | MAGMA first computes a gene-based p value based on the mean association of variants in the gene, accounting for LD between variants; then, competitive gene-set and/or continuous covariate p values are calculated, based on the association of the gene-based p values with the category of interest | https://ctg.cncr.nl/software/magma | de Leeuw et al.39 | 2,231 | 6 | 0.30% | 2.6 |
| FINEMAP + HiC | FINEMAP uses a Bayesian approach to determine which are the most likely causal variants in a locus; candidate causal variants identified by FINEMAP are mapped to genes using chromatin conformation capture HiC data, which represents loops of DNA where regions of the genome interact | http://www.christianbenner.com | Benner et al.31 | 51 | 0 | 0% | 1 |
| Summary-data-based Mendelian randomization (SMR) + heterogeneity in dependent instruments (HEIDI) | integrates summary-level data from GWASs with data from eQTL studies to identify genes whose expression levels are associated with a complex trait because of pleiotropy; the methodology can be interpreted as an analysis to test if the effect size of an SNP on the phenotype is mediated by gene expression; the HEIDI method then uses multiple SNPs in a cis-eQTL region to distinguish pleiotropy from linkage | https://cnsgenomics.com/software/smr/ | Zhu et al.25 | 65 | 0 | 0% | 1 |
| Similarity-based methods | |||||||
| Polygenic priority score (PoPS) | uses gene-level associations computed from GWAS summary statistics to learn joint polygenic enrichments of gene features derived from gene expression, biological pathways, and protein-protein interactions (PPI), assigning a priority score to every protein-coding gene | https://github.com/FinucaneLab/pops | Weeks et al.18 | 486 | 4 | 0.80% | 5.3 |
| Data-driven expression prioritized integration for complex traits (DEPICT) | employs annotated gene sets (including manually curated pathways, molecular pathways from protein-protein interaction screens, and phenotypic gene sets from mouse gene knock-out studies). By calculating, for each gene, the likelihood of membership in each gene set (based on similarities across expression data), 14,461 ‘reconstituted’ gene sets were generated. Using these precomputed gene functions and a set of trait-associated loci, DEPICT assesses whether any of the 14,461 reconstituted gene sets are significantly enriched for genes in the associated loci, and prioritizes genes that share predicted functions with genes from the other associated loci more often than expected by chance. | https://github.com/perslab/depict | Pers et al.17 | 252 | 1 | 0.40% | 3.2 |