Unsupervised learning of cross-modal mappings in multi-omics data for survival stratification of gastric cancer

• Published in: Future oncology (London, England) Vol. 18; no. 2; pp. 215 - 230
• Main Authors: Xu, Jianmin, Yao, Yueping, Xu, Binghua, Li, Yipeng, Su, Zhijian
• Format: Journal Article
• Language: English
• Published: England Future Medicine Ltd 01.01.2022
• Subjects: Aged; BiDNNs; Biomarkers, Tumor - genetics; Cluster Analysis; cross-modal mappings; Datasets as Topic; disease modeling; DNA Methylation; Epigenomics - methods; Female; Follow-Up Studies; gastric cancer; Humans; Kaplan-Meier Estimate; machine learning; Male; Middle Aged; multi-omics integration; patient stratification; risk assessment; Risk Assessment - methods; RNA-Seq - methods; Stomach Neoplasms - genetics; Stomach Neoplasms - mortality; Unsupervised Machine Learning; BiDNNs; risk assessment; gastric cancer; disease modeling; multi-omics integration; patient stratification; cross-modal mappings; machine learning
• ISSN: 1479-6694; 1744-8301
• DOI: 10.2217/fon-2021-1059

This study presents a survival stratification model based on multi-omics integration using bidirectional deep neural networks (BiDNNs) in gastric cancer. Based on the survival-related representation features yielded by BiDNNs through integrating transcriptomics and epigenomics data, K-means clustering analysis was performed to cluster tumor samples into different survival subgroups. The BiDNNs-based model was validated using tenfold cross-validation and in two independent confirmation cohorts. Using the BiDNNs-based survival stratification model, patients were grouped into two survival subgroups with log-rank p-value = 9.05E-05. The subgroups classification was robustly validated in tenfold cross-validation (C-index = 0.65 ± 0.02) and in two confirmation cohorts (E-GEOD-26253, C-index = 0.609; E-GEOD-62254, C-index = 0.706). We propose and validate a robust and stable BiDNN-based survival stratification model in gastric cancer.