Fig. 6

A gastric cancer screening model based on microbial-derived tongue-coating proteins. A Construction workflow of the gastric cancer screening model based on microbial-derived tongue-coating protein. B The t-distributed stochastic neighbor embedding (t-SNE) visualization of gastric cancer and nongastric cancer samples in the training dataset based on the top 50 tongue-coating microbial proteins selected by the GBM model. C The performance of the receiver operating characteristic curve (ROC) to classify patients with gastric cancer and healthy individuals using the 50 microbial protein features in the training data, test data, and multi-center cohort. D The distribution ratio and correspondence between different cohorts and samples. E Accuracy, sensitivity, and specificity of our screening model in different datasets. F Functional enrichment analysis of 50 microbial-derived proteins included in the screening model