Gene expression data can serve for analyzing the genes with changed expressions, the correlation between genes and the influence of different circumstances on gene activities. However, labeling a large number of gene expression data is laborious and time-consuming. The insufficient labeled data pose a challenge to construct the deep learning model. We put forward a novel semi-supervised graph neural network model, Semi-Supervised Feature Weighted Neural Network (SFWN). Firstly, we use the external knowledge of gene expression data for constructing a feature graph, a similarity kernel, and a sample graph for the first time. Later, SGA, a novel semi-supervised learning algorithm, is proposed to extract the data relationship and obtain the global sample structure better. SGCN, a graph sparse module, is also proposed to process sparse representation with gene expression data classification. To overcome the over smoothing problem, a new feature calculation method based on two spaces is proposed to feature representation analysis and calculation in this model. The proposed SFWN model has strong gene expression feature learning and representation ability, and may provide a new insight and tool for relevant disease diagnosis and clinic practice.
The research results, titled “SFWN: A Novel Semi-Supervised Feature Weighted Neural Network for Gene Data Feature Learning and Mining with Graph Modeling”, were published in the well-known international journal of biomedical and health informatics, IEEE Journal of Biomedical and Health Informatics, with the College of Photonic and Electronic Engineering of FNU as the first unit, Wang Qing, a postgraduate, as the first author, and Prof. Chen Guannan as the corresponding author. The work was supported by such projects as the Fujian Provincial Health Education Joint Research Project and Special Projects of the Central Government in Guidance of Local Science and Technology Development.
(Translated by Cai Yuan/ Reviewed by Xie Xiujuan)