1) SNP-metabolite associations (mGWAS)
As of December 2021, 65 mGWAS papers had been found after searching PubMed, Web of
Science, bioRxiv, and medRxiv, after a thorough literature research. The summary statistics were
either collected from publicly available databases or supplementary data from the original
publications. Study-specific significant thresholds were used to pre-filter statistical associations
between metabolites and SNPs. Details of the curated mGWAS dataset can be found in our
Resources page: mGWAS-Explorer
2) SNP to gene mapping
Three options are provided for SNP to gene mapping, including HaploReg, PhenoScanner,
and VEP by using the Application Programming Interface (API) service of each database.
3) LD proxy search
mGWAS-Explorer allow users to search for metabolites/diseases associations with proxies for
SNPs of interest using the HaploReg API or PhenoScanner API. The LD information is based on
the 1000 Genomes Project.
4) SNP-disease association
DisGeNET and OpenGWAS database were used to obtain SNP-disease associations. Particularly, DisGeNET contains both curated and literature data. The curated data include SNP-disease associations from UniProt, ClinVar, GWAS Catalog, and GWASdb. In contrast, OpenGWAS is a database of 244,868,980,310 genetic associations from 42,334 GWAS summary datasets.
Note, OpenGWAS is used to extract outcome GWAS in MR analysis.
5) Gene-metabolite association
Knowledge-based gene-metabolite association information was curated using KEGG, Recon3D,
and the Transporter Classification Database (TCDB).
6) Protein-protein interaction
Information on protein-protein interaction was taken from four well-established PPI databases,
including InnateDB, STRING, HuRI, and Rolland et al.
7) Gene-disease association
DisGeNET was also used to obtain gene-disease associations.
8) Metabolite-disease association
HMDB was used to retrieve metabolite-disease associations.