Several Data Mining Techniques Applied Into Predicting Signal Peptide

Because of its importance in directing protein secretion, signal peptide (SP) is now a hot spot in Bioinformatics. It is really time-consuming and labor-consuming to predict signal peptide and their cleavage sites using experimental methods. Lots of pattern recognition methods have been developed since 1980s and working efficiency has increased largely. The emphases here are the applications of several data mining techniques in the following hot fields of Bioinformatics: applying global alignment algorithms into measuring the similarity between signal sequences, and dealing with the problem of imbalanced training samples caused by sliding window.Main work of this paper includes:1. Global alignment algorithms applied into measuring the similarity between signal sequencesPrediction of signal peptide usually includes two aspects of work: discrimination of a signal peptide and finding the location of their cleavage sites. Although various methods for identifying signal peptides...