Prediction of lysine propionylation sites using biased SVM and incorporating four different sequence features into Chou's PseAAC.

摘要: Abstract Lysine propionylation is an important and common protein acylation modification in both prokaryotes eukaryotes. To better understand the molecular mechanism of propionylation, it to identify propionylated substrates their corresponding sites accurately. In this study, a novel bioinformatics tool named PropPred developed predict by using multiple feature extraction biased support vector machine. On one hand, various features are incorporated, including amino acid composition, factors, binary encoding, composition k-spaced pairs. And F-score method incremental selection algorithm adopted remove redundant features. other machine used handle imbalanced problem training dataset. As illustrated 10-fold cross-validation, performance achieves satisfactory with Sensitivity 70.03%, Specificity 75.61%, accuracy 75.02% Matthew’s correlation coefficient 0.3085. Feature analysis shows that some factors play most roles prediction sites. These results might provide clues for understanding mechanisms propionylation. A user-friendly web-server established at 123.206.31.171/PropPred/.