Isoelectric point optimization using peptide descriptors and support vector machines

• Published in: Journal of proteomics 2012-04, Vol.75 (7), p.2269-2274
• Main Authors: Perez-Riverol, Yasset, Audain, Enrique, Millan, Aleli, Ramos, Yassel, Sanchez, Aniel, Vizcaíno, Juan Antonio, Wang, Rui, Müller, Markus, Machado, Yoan J., Betancourt, Lazaro H., González, Luis J., Padrón, Gabriel, Besada, Vladimir
• Format: Article
• Language: English
• Subjects: algorithms; amino acid sequences; amino acids; Biological and medical sciences; Diverse techniques; Fundamental and applied biological sciences. Psychology; Isoelectric point; Molecular and cellular biology; Peptide descriptors; peptides; proteomics; Support vector machine
• ISSN: 1874-3919
• DOI: 10.1016/j.jprot.2012.01.029

IPG (Immobilized pH Gradient) based separations are frequently used as the first step in shotgun proteomics methods; it yields an increase in both the dynamic range and resolution of peptide separation prior to the LC-MS analysis. Experimental isoelectric point (pI) values can improve peptide identifications in conjunction with MS/MS information. Thus, accurate estimation of the pI value based on the amino acid sequence becomes critical to perform these kinds of experiments. Nowadays, pI is commonly predicted using the charge-state model [1], and/or the cofactor algorithm [2]. However, none of these methods is capable of calculating the pI value for basic peptides accurately. In this manuscript, we present an new approach that can significant improve the pI estimation, by using Support Vector Machines (SVM) [3], an experimental amino acid descriptor taken from the AAIndex database [4] and the isoelectric point predicted by the charge-state model. Our results have shown a strong correlation (R2=0.98) between the predicted and observed values, with a standard deviation of 0.32 pH units across the complete pH range. [Display omitted] ► First time that the support vector machines is used to predict the isoelectric point. ► Current isoelectric point algorithm has been compare for the same experiment. ► SVM algorithm shows a better correlation and standard deviation. ► Current model could be used to reduce the number of false positive identification.