- 作者: 樑桂兆, 舒茂, 李聲時
- 作者服務機構: College of Bioengineering, Chongqing University, Chongqing 400030, P. R. China
- 中文摘要: A new set of amino acid descriptors, SZOTT (Scores of Zero dimension, One dimension, Two dimension, Three dimension) was derived from principal component analysis on a matrix of 1369 descriptors including 0D, 1D, 2D and 3D information of 20 coded amino acids. SZOTT scales were then employed to represent structures of HLA-A*0201 restrictive CTL epitopes, and further, quantitative sequence-activity modelings (QSAMs) were constructed based on partial least squares and back-propagation artificial neural networks. Satisfying results show that structural information related to biological activities of HLA-A*0201 restrictive CTL epitopes can be preferably represented by SZOTT scales, which may be a useful structural representation technique for study of QSAMs of peptides.
- 英文摘要: --
- 中文關鍵字: Amino acids; CTL epitopes; SZOTT; QSAM; Principal component analysis; Partial least squares; Back-propagation artificial neural networks.
- 英文關鍵字: --