Inferring Disease-Related Domain Using Network-Based Method

Domain-domain interaction (DDI) network analysis has been widely applied in the investigation of the mechanisms of diseases. This project focus on mapping the melanoma-related mutations to domain and domain-domain interaction (DDI) level and use potential correlation method to find the human domains that have never been found to have anything to do with the melanoma-related study. Firstly, we extract melanoma-related mutations from COSMIC database, then map the mutations to protein database UniProt and domain database Pfam to find the melanoma-related domains; Secondly, we get the melanoma-related DDI information and the human DDI information from the DDI database iPfam; Thirdly, we construct the melanoma-related DDI network and human DDI network and then combine two approaches to use potential correlation method to analyze the two DDI networks. Finally, we find that among all the human domains who have potential correlation relationship with melanoma, there are 27 human domains that have never been found to have anything to do with melanoma in the existing literature and study. The result shows the effectiveness of our method and further study based on these human domains may lead to some new methods to cure melanoma.