SPAAN: a software program for prediction of adhesins and adhesin-like proteins using neural networks

• DOI: 10.1093/bioinformatics/bti028
• Publication Date: 2004-09-17T00:13:37Z
• Authors: Gaurav Sachdeva; Kaushal Kumar; Preti Jain; Srinivasan Ramach...
• Keywords: 0303 health sciences; Models, Statistical; Sequence Homology, Amino Acid; Models, Biological; 3. Good health; Structure-Activity Relationship; 03 medical and health sciences; Models, Chemical; Sequence Analysis, Protein; Animals; Humans; Protein Isoforms; Neural Networks, Computer; Adhesins, Bacterial; Sequence Alignment; Algorithms; Software

Abstract Motivation: The adhesion of microbial pathogens to host cells is mediated by adhesins. Experimental methods used for characterizing adhesins are time-consuming and demand large resources. availability specialized software can rapidly aid experimenters in simplifying this problem. We have employed 105 compositional properties artificial neural networks develop SPAAN, which predicts the probability a protein being an adhesin (Pad). Results: SPAAN had optimal sensitivity 89% specificity 100% on defined test set could identify 97.4% known at high Pad value from wide range bacteria. Furthermore, facilitated improved annotation several proteins as Novel were identified 17 pathogenic organisms causing diseases humans plants. In severe acute respiratory syndrome (SARS) associated human corona virus, spike glycoprotein nsps (nsp2, nsp5, nsp6 nsp7) having adhesin-like characteristics. These results offer new lead rapid experimental testing. Availability: freely available through ftp://203.195.151.45 Contact: [e-mail]