Abstract
Semantic similarity measures using Gene Ontology (GO) assess the semantic similarity between two proteins on the basis of their respective GO annotations, which represent specific features of the proteins. Proteins will interact with greater likelihood if they have greater number of similar GO annotations. In this paper, we design a semantic similarity measure by combining topological features of the GO graph with information contents of the GO terms. Here, we perform a clustering of the GO graph and determine membership of GO terms to the cluster centers based on their respective shortest path lengths. We compare the performance of our semantic similarity measure with other prominent semantic similarity measures based on correlation with sequence similarity and Pfam similarity using Homo sapiens protein-protein interaction dataset.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have