Cancer typing information processing method based on multi-omic data

Abstract

The invention belongs to the field of biology and medical gene technology, and discloses a cancer typing information processing method based on multi-omic data. The method comprises the steps of preprocessing single-base mutation data of the multi-omic data for converting to a three-base mutation mode; performing non-negative matrix decomposition on the three-base mutation matrix which is obtainedthrough preprocessing, performing k-means cluster analysis on the decomposed characteristic matrix, selecting an optimal k by means of a contour coefficient, and obtaining a subtype label of a cancersample in guidance of mutation; afterwards preprocessing the residual omic data, performing integrated cluster analysis by means of an iclusterplus tool, determining a cluster model according to a minimum Bayes criterion, performing subtype definition on the sample again, then combining the results in two times of typing by means of a customized algorithm, and determining a final cancer subtype label of the sample. The cancer typing information processing method has advantages of convenient data acquisition, convenient tool operation and relatively high result reliability.

Description

technical field [0001] The invention belongs to the field of biology and medical gene technology, and in particular relates to a method for processing cancer typing information based on multi-omics data. Background technique [0002] With the improvement of gene sequencing technology, researchers can obtain the molecular level changes of deoxyribose nucleic acid (Deoxyribonucleic acid, DNA) in the sequenced samples, confirming that the same organ may have different DNA molecular structures in different samples, while different organs may be different. The existence of the same DNA molecular structure broke the long-standing traditional cognition of typing. People have reason to speculate that the molecular structure of patients with different cancer types, different cancer subtypes of the same cancer type, etc. is more complicated, and solving these difficult problems requires more time and cost investment for large-scale genetic data analysis. [0003] In February 2018, th...

AI Technical Summary

Problems solved by technology

The shortcoming of this method is that the multi-omics data are directly fused together for analysis, without considering the different degrees of influence of cancer typing among multiple omics data

[0012] (2) Based on the different degrees of influence of multi-omics data on cancer typing, hierarchical analysis is required, so it is also a challenge to fuse the results of cancer typing at different levels, that is, how to combine the results of each type of omics data Largest unification of cancer subtype labels

Images