Classification of Antimicrobial Peptides by Using the p-spectrum Kernel and Support Vector Machines Paola Rond´ on-Villarreal, Daniel A. Sierra, and Rodrigo Torres Universidad Industrial de Santander, Carrera 27 calle 9, Bucaramanga, Colombia paitorv@gmail.com, dasierra@uis.edu.co, rtorres@uis.edu.co http://www.uis.edu.co Abstract. In the last decades, antibiotic resistance of pathogenic mi- croorganisms constitutes a great problem of public health at global level. Multidrug-resistant bacteria cannot be controlled with the existing medi- cations causing thousands of deaths every year. In the fight against these bacteria, antimicrobial peptides have appeared as a promising solution as therapeutic agents against pathogens. For this reason, rational design of these chemical compounds have been explored by the scientific commu- nity in order to achieve significant improvements that could lead to the discovery of new antibacterial medicine. In this sense, the present work proposes the use of the p-spectrum kernel with support vector machines to classify antimicrobial peptides, thus considering only the information of the order of the amino acids inside the peptide sequences. The results were satisfactory and suggest that this information should be considered in the rational design of antimicrobial peptides. Keywords: antimicrobial peptides, kernel methods, support vector machines. 1 Introduction Nowadays, there are multiple microorganisms that are becoming resistant to the existing medications. Among them, the multidrug-resistant bacteria kill thou- sands of people across the globe every year, which represents an extreme risk for the humanity. Only in the United States the situation is becoming critical. More than 40 states have been affected with at least one patient infected with CRE (carbapenem-resistant Enterobacteriaceae) bacteria. The picture is disturbing considering that current medications can no control the infections caused by these super bacteria. On the other hand, each year in the United Kingdom die about 2,500 patients by bloodstream infections caused by multidrug resistant bacteria. The major concern is that the available medication can no kill these organisms and the pharmaceutical industry is not so much interested in antimicrobial medicine developments. L.F. Castillo et al. (eds.), Advances in Computational Biology, 155 Advances in Intelligent Systems and Computing 232, DOI: 10.1007/978-3-319-01568-2_ 23, c  Springer International Publishing Switzerland 2014