Progressively Helical Multi-omics Data Fusion GCN and Its Application in Lung Adenocarcinoma

Compared to single-omics data, utilizing multi-omics data helps gain a more comprehensive understanding of the occurrence and development cancer, which emphasizes necessity developing efficient fusion approaches. In this study, novel framework based on graph convolution neural networks with progressively helical strategy, named phMFGCN, is proposed effectively integrate multiple omics data. To demonstrate effectiveness our in addressing challenges multi fusion, phMFGCN other widely-used machine learning methods conducted comparative experiments predicting gene-gene interactions lung adenocarcinoma. The results illustrated that outperforms models an accuracy 97.94%. Additionally, 506 new predicted by have been validated databases such as BioGrid. Finally, it was used perform gene function prediction, were inconsistent existing research, for examples: Sam68, DHX9, HNRNPK involved regulating adenocarcinoma related pathways simultaneously. All these universality different clinical tasks can provide reference target genes or cancer mechanism research treatment practice.