Rapid feature selection method based on whole genome sequence SNP data

Abstract

The invention relates to the technical field of SNP feature selection, and discloses a rapid feature selection method based on whole genome sequence SNP data, which comprises the following steps: (A) preprocessing the whole genome sequence data to obtain all SNP sites; (B) filtering irrelevant SNP features by adopting an FFS algorithm, and reserving required SNP features; (C) performing key degree sorting on the SNP features reserved in the step (B); (D) screening out significant SNPs according to a key degree sorting result. According to the method, the significant SNPs are screened out from all SNPs through SNP calling, rapid feature selection and key degree sorting, and the method has the advantages of being high in accuracy and high in high-dimensional sample feature selection efficiency and can be used for positioning pathogenic genes, drug-resistant genes and the like.

Description

technical field [0001] The invention relates to the technical field of SNP feature selection, in particular to a rapid feature selection method based on whole genome sequence SNP data. Background technique [0002] In the field of data mining, feature selection is the process of screening out the minimum feature subset for identifying targets from the original feature set. Feature selection for data with high-dimensional and small-sample characteristics is one of the research hotspots in this field. Feature selection can reduce the dimensionality of the feature space, thereby speeding up the overall calculation speed of the model; screening out redundant "noise" data; increasing the classification accuracy of the classification model; and helping to reveal the information contained in the data. Commonly used data analysis methods have the characteristics of sample tendency, and the efficiency and accuracy of sample data analysis are low. [0003] Single nucleotide polymorp...

AI Technical Summary

Problems solved by technology

However, the current SNP feature selection methods usually have problems such as low accuracy and low efficiency of high-dimensional sample feature selection.

For example, the Chinese patent document with the announcement number CN102629305B discloses a feature selection method for SNP data, which improves and combines the Relief algorithm and the SVM-RFE algorithm to perform feature selection on SNP data. Compared with the traditional single Relief As far as the algorithm and the SVM-RFE algorithm are concerned, the analysis efficiency and accuracy have been improved, but there are still the following problems: the combination of the two algorithms only synthesizes their respective advantages, and the existing problems have not been solved, so the analysis efficiency and a limited increase in accuracy

Images