ANCO-GeneDB: annotations and comprehensive analysis of candidate genes for alcohol, nicotine, cocaine and opioid dependence

Abstract

Studies have shown that genetic factors play an important role in the risk to substance addiction and abuse. So far, various genetic and genomic studies have reported the related evidence. These rich, but highly heterogeneous, data provide us an unprecedented opportunity to systematically collect, curate and assess the genetic and genomic signals from published studies and to perform a comprehensive analysis of their features, functional roles and druggability. Such genetic data resources have been made available for other disease or phenotypes but not for major substance dependence yet. Here, we report comprehensive data collection and secondary analyses of four phenotypes of dependence: alcohol dependence, nicotine dependence, cocaine dependence and opioid dependence, collectively named as Alcohol, Nicotine, Cocaine and Opioid (ANCO) dependence. We built the ANCO-GeneDB, an ANCO-dependence-associated gene resource database. ANCO-GeneDB includes resources from genome-wide association studies and candidate gene-based studies, transcriptomic studies, methylation studies, literature mining and drug-target data, as well as the derived data such as spatial-temporal gene expression, promoters, enhancers and expression quantitative trait loci. All associated genes and genetic variants are well annotated by using the collected evidence. Based on the collected data, we performed integrative, secondary analyses to prioritize genes, pathways, eQTLs and tissues that are significantly enriched in ANCO-related phenotypes.

Figure 1

Schematic diagram describing data collection, features and applications of ANCO-GeneDB. The data in ANCO-GeneDB were collected and curated from various sources and were organized at three levels: (i) genes and genetic variants with direct association evidence in the ANCO phenotypes; (ii) indirect annotations on ANCO genes and genetic variants; and (iii) secondary analyses, including TSEA, Sherlock, PASCAL and MetaXcan analyses. Full names of the abbreviations are provided in the main text.

Figure 2

Comparison of genetic variants and genes. (A) Venn diagram comparing the genetic variants associated with each phenotype. (B) Venn diagram comparing ANCO genes (all genes obtained from two approaches). (C) Venn diagram comparing SNP-mapped genes. (D) Venn diagram comparing genes from two sources (SNP-mapped genes and genes extracted from the literature).

Figure 3

Secondary analysis of ANCO-associated genetic variants and genes. (A) TSEA of ANCO-associated genes (each individual gene set and the combined) using 47 GTEx (v7) tissues. Color is proportional to –log10 (P-value), where P-value was obtained from Fisher’s Exact Test. For each gene set, the top 3 most significantly enriched tissues are labeled. (B) eQTLEnrich analysis of the GWAS summary statistics for AD and ND (GTEx v6). The x-axis shows the adjusted fold enrichment. The black lines display 95% confidence interval. (C) Manhattan plots showing the distribution of gene-based P-values obtained by PASCAL analysis. The blue line indicates P = 1 × 10⁻⁵ and red line for P = 2.31 × 10⁻⁶ (Bonferroni correction threshold). (D) Distribution of the pathway enrichment analysis by PASCAL. The green line indicates P = 0.05 and red line for P = 4.64 × 10⁻⁵ (Bonferroni correction threshold).

Figure 4

The SNP Map page. A genome-wide display of all the collected SNPs for ANCO phenotypes. Each node indicates an SNP and node color represents the phenotype. The user can click each SNP node linking to the detailed SNP information page and filter out the data based on P-value cut-off or phenotype (Note: the display of SNPs on chromosomes is adapted from the GWAS Catalog).