Gene inverse mutation detecting method based on feature mining

Abstract

The invention relates to a gene inverse mutation detecting method based on feature mining. The method comprises the steps of acquiring a gene inverse mutation candidate sample set; performing featureextraction on the gene inverse mutation candidate sample set, and obtaining an overturning variation feature set; performing feature selection on the overturning variation feature set, and obtaining an inverse variation effective feature set; performing validity testing on the inverse variation feature set; training an SVM classifier by means of the inverse variation feature set which is extractedfrom simulating samples, and establishing an SVM classifier model; and generalizing true data for obtaining a true inverse mutation data set. Compared with a routine detecting method, the gene inverse mutation detecting method has advantages of effectively improving recall rate and accuracy in detecting the inverse mutation and improving inverse mutation detecting validity through sufficiently mining the inverse mutation characteristic and utilizing the SVM classifier of a machine learning model.

Description

technical field [0001] The invention relates to a gene reversal variation detection method, in particular to a feature mining-based gene reversal variation detection method. Background technique [0002] Variations in the human genome are closely related to human evolution and disease risk. Genome structural variation is a kind of human genome variation, which refers to sequence changes and positional relationship changes on the genome. Structural mutations in genes can lead to abnormalities in life performance, and once they occur, they often have a major impact on living organisms, such as birth defects and cancer. Gene flipping variation is a relatively complex structural variation. The current methods for detecting flipping variation are mainly gene flipping mutation detection methods based on next-generation sequencing data, including paired-end mapping analysis, sequencing fragment split alignment, and sequence splicing. These three methods have achieved certain resu...

AI Technical Summary

Problems solved by technology

Gene flipping variation is a relatively complex structural variation. The current methods for detecting flipping variation are mainly gene flipping mutation detection methods based on next-generation sequencing data, including paired-end mapping analysis, sequencing fragment split alignment, and sequence splicing. These three methods have achieved certain results in the detection of structural variation, but because the flipping mutation is a more complex structural variation, the existing tools for detecting flipping mutations only use one or two conventional detection strategies, and the conventional strategy is only Based on the sequence itself without mining and using more flip mutation features to deal with the uniqueness and complexity of flip mutations, resulting in large limitations in the recall and precision of detecting flip mutations

Images