Deep learning method for predicting prognosis risk of cancer patient based on multi-omics data

Abstract

The invention discloses a deep learning method for predicting the prognosis risk of a cancer patient based on multi-omics data, which is used for predicting the prognosis risk of the cancer patient and comprises the following steps: S1, acquiring clinical data Y of a target cancer patient and corresponding multi-omics expression data X from an existing public data set; S2, constructing a deep neural network; S3, updating the weight theta of the cancer multi-omics data Xp and the clinical information Yp of the patient of the existing common data set through the constructed deep neural network to obtain a pre-training network Np based on the common data set; S4, training the network Np again until the training frequency epoch reaches the operation upper limit, thereby obtaining a risk prediction network Nf; and S5, selecting the first n gene features of the Importance coefficient of the target cancer patient by using an XGboost algorithm, and improving the risk prediction network Nf to obtain a final risk prediction model. According to the method, the robustness of the prediction model is improved, and the prognosis risk of the cancer patient is predicted more accurately by using multi-omics data.

Description

technical field [0001] The present invention relates to the technical field of cancer patient survival analysis, and more specifically, relates to a deep learning method for predicting the prognosis risk of cancer patients based on multi-omics data. Background technique [0002] In recent years, the high incidence of cancer has promoted the development of medical auxiliary technology. Prognostic risk analysis is a key medical auxiliary technology, which can assist in the selection of different treatment options according to the potential risk of prognosis of different patients. [0003] Most of the methods for predicting cancer prognosis are realized by analyzing the expression data of a single group, such as using gene mRNA expression data, methylation data, or miRNA data, etc. Moreover, there are strong complementary effects and interactions between molecules at different levels, so the results of single-omics data analysis often can only provide one-sided information. In...

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a deep learning method for predicting the prognosis risk of cancer patients based on multi-omics data in order to overcome the above-mentioned defects that the accuracy rate of prognosis risk prediction is not high and the target data set cannot be solved.

Images