The Research Of Disease Phenotype Similarity Mining Algorithm In OMIM Text

The purpose of biomedical text mining is using the text mining technologies to help biomedical researchers find the information needed, and discover the hidden biomedical knowledge from massive literatures more effectively.Make biomedical text mining on Online Mendelian Inheritance in Man (OMIM) to get the similarity of the described phenotype in OMIM records. Computing and analyzing the similarity is the foundation of the following study including projections of related diseases, has its very important scientific significance.This article makes text mining on OMIM, which contains human disease phenotypes, to discover the similarity of phenotype, have the function related genes, which reflects the biological module interaction. There are two methods for the disease phenotype similarity in this article. They are separately based on the vector space and based on the scene pattern. By the two methods we can obtain the similarity of disease phenotype in OMIM text. This article takes word stemming of the key verb which appears in the OMIM text to enhance the precision. And it uses the synonym collection (Synset) of WordNet to refine the similarity. It proposed a method; Compute the similarity of the description word of OMIM entity to optimize the disease phenotype similarity obtained based WordNet.Applied the similarity of disease phenotype to disease gene prediction, we can get that the method based on scene pattern is better than based on vector space.