Method for utilizing machine learning to predict complex disease susceptibility locus

Abstract

The invention discloses a method for utilizing machine learning to predict a complex disease susceptibility locus. The method comprises the following steps of 1, collecting a known complex disease susceptibility locus as a positive set of a machine learning model, predicting a locus irrelevant to a complex disease according to the positive set as a negative set, and annotating an epigenetic regulation element; 2, utilizing machine learning to establish a complex disease epigenetic model; 3, predicting all loci in a whole-genome range according to the established model to obtain a final prediction result as a potential susceptibility locus of the complex disease. According to the method for utilizing machine learning to predict the complex disease susceptibility locus, epigenetic information and genome DNA information are combined, epigenetic element features are extracted through machine learning, the susceptibility locus of the complex disease is further predicted in a whole-genome range, heritability explained by the found susceptibility locus can be obviously improved, and a potential target is provided for subsequent medicine design and disease detection.

Description

technical field [0001] The invention relates to the technical field of prediction of complex disease susceptibility loci, in particular to a screening method for predicting complex disease susceptibility loci using machine learning. Background technique [0002] In recent years, genome-wide association analysis has become the hottest and most effective research method for revealing complex disease susceptibility loci (Single nucleotide polymorphism, SNP). Using this method, more than 2,000 papers have been published in international high-level journals, and nearly 10,000 complex disease susceptibility loci have been successfully identified. Although the results of genome-wide association analysis are quite fruitful, it is far from what scientists expected - to find most of the disease susceptibility loci. For a specific complex disease, the reported disease susceptibility loci accumulatively explain less than 15% of the genetic variation of the disease, and there are still ...

AI Technical Summary

Problems solved by technology

At present, there are multiple databases revealing genomic epigenetic information, but hundreds of millions of genetic markers and multi-dimensional element information have brought great challenges to the prediction of genetic loci

Images