Fuzzy pattern extraction for classification of protein sequences

Text mining is an important research area in applied statistics. The present article addresses an important problem from the Bioinformatics field, viz. classification of protein sequences as soluble proteins and inclusion body forming proteins when over-expressed in Escherichia coli (E. coli), using text mining and machine learning techniques. We propose a text mining based algorithm to extract patterns from the protein sequences that are later used in support vector classification algorithm. We report the best classification results for this dataset compared to the existing state of the art. Our algorithm is quite general and can be applied to any biological text data. The extracted patterns may give further insight in underlying dynamics of the sequences that decide the corresponding class membership.