Application of fuzzy clustering analysis to compound datasets for drug lead identification

摘要: Recently, the increasing number of chemical compound datasets to be screened has been growing rapidly due fast developments high-throughput screening in drug discovery. These requires selection methods which have become one main technique discovery especially lead identification process. Thus, finding best method is needed pharmaceutical industry ensure accurate results this One most used cluster-based selection, involves subdividing a set compounds into clusters and choosing or small from each cluster. In non-overlapping such as Ward's, Group Average, Jarvis Patrick's K-means are preferred cluster diverse compounds. However, there little study on overlapping fuzzy c-mean (FCM) c-varieties (FCV) clustering algorithms. Therefore, these two algorithms applied their performance compared based effectiveness terms separation between actives inactives (Pa) different mean intercluster molecular dissimilarity (MIMDS). The analysis shows FCM gives compare FCV Pa indicating that promising use But, perform better than term MIMDS when higher fuzziness index value concerned.