This application provides a method,
system, and application for identifying potential
drug targets, relating to the field of
bioinformatics. The method includes: S1, constructing genetic instrumental variables related to a predetermined
disease based on a multidimensional dataset, wherein the multidimensional dataset includes: a
disease genome dataset, a dataset of druggable genes, and a dataset of
plasma protein quantitative trait loci; S2, analyzing the genetic instrumental variables using Mendelian
randomization to determine the potential causal relationship between genes and the predetermined
disease; S3, validating the potential causal relationship between genes and the predetermined disease using a
drug-
gene dataset to identify potential
drug targets for the predetermined disease. The aforementioned method, by comprehensively integrating disease
genome, druggable
gene, and
protein quantitative trait locus data, effectively eliminates the interference of horizontal pleiotropic effects on the identification of potential drug targets and eliminates the
confusion that may be caused by
linkage disequilibrium, thereby improving the accuracy and reliability of potential
drug target screening.