Brain metastasis prediction method and system based on DNA characteristics and storage medium

Abstract

The invention discloses a brain metastasis prediction method and system based on DNA features and a storage medium, and belongs to the technical field of biomedicine. The method comprises the following steps: S1, identifying an initial differential methylation CpG site; s2, selecting differential methylation CpG sites related to the brain metastasis time from the initial differential methylation CpG sites; s3, randomly dividing the sample data into a discovery set and a verification set, and constructing a methylation risk scoring model by using the discovery set; s4, evaluating the prediction capability of the methylation risk scoring model by using the verification set, and training the methylation risk scoring model according to an evaluation result until the prediction capability reaches a preset requirement; and S5, inputting the plasma cfDNA data into the trained methylation risk scoring model to generate a comprehensive risk assessment map, and predicting the brain metastasis risk probability. Based on DNA methylation characteristics, higher prediction accuracy is provided, a high-risk patient can be recognized earlier, and through early prediction of the risk of brain metastasis, intervention is performed in advance, and the life quality of the patient after healing is improved.

Images