Multi-omics cancer data integrating and analyzing method based on similarity fusion

Abstract

The invention discloses a multi-omics cancer data integrating and analyzing method based on similarity fusion. The method comprises the steps of calculating a local similarity network, fusing multiplelocal similarity networks, typing according to a global similarity network, and backtracking the characteristic in an original data source according to the global similarity network. Compared with the prior art, the method is advantageous in that through modeling a similarity network connecting path which gradually advances, a fusion algorithm of multiple similarity networks is realized; a more complicated network structure can be described; and higher accuracy and higher stability are realized. Through a consistent alternating multiplier method, quick solving of a network fusion model is realized. The multi-omics cancer data integrating and analyzing method has advantages of utilizing the integrated global similarity network to typing of a cancer patient, obtaining types of the patientswith substantial prognosis difference, and combining with a multi-group characteristic selecting method for screening the key target characteristic. The selected characteristic has a potential of becoming a latent biological marker.

Description

technical field [0001] The present invention relates to the technical field of biological information, in particular to a multi-omics cancer data integration analysis method based on similarity fusion. Background technique [0002] In current clinical medical practice, cancer is often classified and treated according to its tissue origin and pathological features. However, with the development of sequencing technology and human genome research, a large number of studies have shown that the pathological system of tumors at the molecular level can well characterize its occurrence characteristics and development stages. The process of tumorigenesis and development is often accompanied by genomic variation caused by somatic gene mutations, epigenetic changes, individual differences, and environmental influences. Traditional analysis based on single genome data is difficult to capture the heterogeneity of all biological processes. and clearly distinguish the phenotypes. Therefo...

AI Technical Summary

Problems solved by technology

The main problem of the network-based method is that the modeling of the network structure and transmission path is simple and cannot describe the complex network structure of multi-omics data, and it does not provide a method to feed back the fusion results to the original omics data

Therefore, for cancer multi-omics data with high dimensionality, high noise, and various distributions, there is still a lack of efficient, accurate, and robust integration methods for cancer molecular analysis

Images