Protein Subcellular Localization Prediction for Fusarium graminearum∗

The fungal pathogen Fusarium graminearum (telomorph Gibberella zeae) is the causal agent of several destructive crop diseases. Investigating subcellular localizations of F. graminearum proteins can provide insight into pathogenic mechanisms underlying F. graminearum-host interactions. In this paper, we design a novel balanced ensemble classifier based on support vector machines (SVMs) to predict F. graminearum proteins’ subcellular localization from the primary sequence. The method is performed with a fungi dataset collected from UniProtKB database. In addition, we utilize SCL-BLAST (SubCellular Localization BLAST) to transfer annotations of homologous proteins to the target uncharacterized protein. We make three fold contributions to this filed. First, we present a new algorithm to cope with imbalance problem that arises in protein subcellular localization prediction, which can improve prediction accuracy significantly. Second, we employ feature selection techniques to find out most informative features for each compartment, and reduce computation cost and improve prediction accuracy at the same time. Third, we use BLAST to complement SVMs based methods, which makes our prediction more effective.