Label Propagation Based Semi-supervised Feature Selection to Decode Clinical Phenotype of Huntington’s Disease

Abstract

Huntington’s disease is a type of neurodegenerative disease caused by gene HTT. To date, its molecular pathogenesis is still unclear. Clinically, behavior, cognitive, and mental function are affected progressively. With the rapid development of sequencing technologies, it is possible to explore the molecular mechanisms at the genome-wide transcriptomic level using computational methods. Our previous studies have shown that it is difficult to distinguish disease genes from non-disease genes. To understand the molecular pathogenesis under complex clinical phenotypes during the disease progression, it is better to identify biomarkers corresponding to different disease stage. Therefore, in this study, we designed a label propagation based semi-supervised feature selection approach (LPFS) to identify disease-associated genes corresponding to different clinical phenotypes. LPFS selects disease-associated genes corresponding to different disease stage through the alternative iteration of label propagation clustering and feature selection. We then conducted an enrichment analysis to understand gene functions and affected pathways during the disease progression, thus to decode the changes in individual behavioral and mental characteristics during neurodegenerative disease progression at the gene expression level. Our results have shown that LPFS performs better in comparison with the-state-of-art methods. We found that TGF-beta signaling pathway, olfactory transduction, cytokine-cytokine receptor interaction, immune response, and inflammatory response were gradually affected during the disease progression. In addition, we found that the expression of Ccdc33, Capsl, Al662270, and Dlgap5 were seriously changed caused by the development of the disease.