Related Experiment Video
Updated: May 9, 2026
Establishment and Evaluation of a Risk Prediction Model for Pathological Escalation of Gastric Low-Grade Intraepithelial Neoplasia
Published on: February 16, 2024
Classification of gastric cancer subtypes using ICA, MLR and Bayesian network
Hiroyuki Ishii1, Hiroki Sasaki, Kazuhiko Aoyagi
1Department of Bioscience and Bioinformatics, Kyushu Institute of Technology, Fukuoka, Japan.
Abstract:
The purpose of this study is to extract genes that could distinguish between two subtypes (diffuse and intestinal) of gastric cancer (GC) from their gene expression profile, to build a classifier based on the genes and to investigate the relationships among the selected genes, because there are big differences in the survival curve and the medical treatment between the subtypes. First, we applied Wilcoxon test, independent component analysis (ICA) and multiple logistic regression (MLR) to the profile, and conducted a classifier consisting of only three genes. We validated the classifier using test data with new 9 samples (blind test). The classifier yielded the accuracy of 100%. Finally, using the 18 genes selected by MLR and specific to each subtype, Bayesian Network was constructed, and compared with the network from GeneMANIA. Consequently, these networks were very similar each other.
