High-throughput Feature Identification Method for Cell Signaling Pathways Combined with Big Data Analysis

In view of the problem that traditional experimental methods in current cell signaling research are difficult to efficiently process massive data and cannot accurately identify key path features, this paper proposes a high-throughput feature recognition method combined with big data analysis. This method first integrates cell signaling data from multi-source databases, and performs data cleaning, normalization, and feature extraction through a parallel computing framework; then, an improved FPGrowth association rule mining algorithm and convolutional neural network are used to identify key features, and a signal network is constructed using graph theory, and key nodes are determined through centrality analysis. The experiment is verified on KEGG, Reactome and BioGRID data sets, and 15 key signal pathway features are successfully identified, with the highest accuracy increased to 100%, and the average processing time is only 10.7 seconds at a scale of 100,000 records; among them, 5 new features strongly correlated with cancer markers are found in the EGFR pathway of KEGG. This method significantly improves the efficiency and accuracy of signal transduction pathway analysis, and provides a new tool for disease mechanism research and target discovery.