文档介绍：Comparison Study of Peptide Retention Time
Prediction Model Based on Five kinds of Amino Acid
Descriptors in HPLC by Support Vector Machine
Jiajian YIN
Department of chemistry
College of life and science, Sichuan Agricultural University
Yaan, ,625014
yinjj@
Abstract—Based on amino acid descriptors(z-scales, c-scales, physicochemical properties, which making them possible to
ISA-ECI,MS-WHIM and PRIN) and additive method, evaluation theoretically predict the retention time of a peptide from
of predict performance of five amino acid descriptors in peptide knowledge of its amino acid content. The literature[1-4]
QSRR(Quantitative structure-retention relationships) with 101 utilized amino position or quantum chemical
promiscuous peptides in High-Performance Liquid Chromato- parameter of special putation to study quantitative
graphy by support vector regression(SVR) is made in the article, structure-retention relationships (QSRR), but they need large
and RBF(radical basis function) is selected as kernel function. numbers putation and they are time-consuming to
Using leave-one-out cross-validation (LOO-CV), we suppose that acquisition molecular structure information parameter of
predicting accuracy of ISA-ECI is better than the other representation peptide, and not able pute biggish peptide
descriptors in SVR with RBF. The pred