Gene Function Similarity Method Based on Data Fusion

Abstract

At present, gene function similarity calculation is a hot field. The mainstream methods are based on gene ontology to calculate gene function similarity. But the purpose of gene ontology is not to congenital deficiency of gene expression function, but biological pathways have advantages in gene expression function. The single gene ontology information expression ability is weak, a gene function similarity algorithm based on data fusion is proposed, which uses gene ontology and biological pathway fusion. The disease gene identification experiment based on machine learning proves that the proposed algorithm is superior to the method of using gene ontology alone and the method of using biological pathways alone. The precision value P reaches 98.9%, and the minimum recall rate R reaches 97.7%. It shows that the fusion of gene ontology and biological pathway data can effectively improve the accuracy of gene function similarity.