A rule sets ensemble for predicting MHC II-Binding peptides

Computational modeling of predicting which peptides can bind to a specific MHC molecule is necessary for minimizing the number of peptides required to synthesize and advancing the understanding for the immune response. Most prediction methods hardly acquire understandable knowledge and there is still some space for the improvements of prediction accuracy. Thereupon, Rule Sets Ensemble (RSEN) algorithm based on rough set theory, which utilizes expert knowledge of bindingimotifs and diverse attribute reduction algorithms, is proposed to acquire understandable rules along with the improvements of prediction accuracy. Finally, the RSEN algorithm is applied to predict the peptides that bind to HLA-DR4(B1*0401). Experimentation results show: 1) compared with the individual rule sets, the rule sets ensembles have significant reduction in prediction error rate; 2) in prediction accuracy and understandability, rule sets ensembles are better than the Back-Propagation Neural Networks (BPNN).

[1]  J. Hammer,et al.  New methods to predict MHC-binding sequences within protein antigens. , 1995, Current opinion in immunology.

[2]  H. Rammensee,et al.  SYFPEITHI: database for MHC ligands and peptide motifs , 1999, Immunogenetics.

[3]  Vladimir Brusic,et al.  Prediction of MHC class II-binding peptides using an evolutionary algorithm and artificial neural network , 1998, Bioinform..

[4]  D. Madden.  The three-dimensional structure of peptide-MHC complexes. , 1995, Annual review of immunology.

[5]  David J. Miller,et al.  Critic-driven ensemble classification , 1999, IEEE Trans. Signal Process..

[6]  Kun Yu,et al.  Methods for Prediction of Peptide Binding to MHC Molecules: A Comparative Study , 2002, Molecular medicine.

[7]  S. Stevanović,et al.  Combining computer algorithms with experimental approaches permits the rapid and accurate identification of T cell epitopes from defined antigens. , 2001, Journal of immunological methods.

[8]  Jun Zeng,et al.  Predicting sequences and structures of MHC-binding peptides: a computational combinatorial approach , 2001, J. Comput. Aided Mol. Des..

[9]  Jerzy W. Grzymala-Busse,et al.  Rough Sets , 1995, Commun. ACM.

[10]  Joachim Diederich,et al.  Survey and critique of techniques for extracting rules from trained artificial neural networks , 1995, Knowl. Based Syst..

[11]  William S. Lane,et al.  Predominant naturally processed peptides bound to HLA-DR1 are derived from MHC-related molecules and are heterogeneous in size , 1992, Nature.

[12]  Wang Ju,et al.  Reduction algorithms based on discernibility matrix: The ordered attributes method , 2001, Journal of Computer Science and Technology.

[13]  L Raddrizzani,et al.  Different modes of peptide interaction enable HLA-DQ and HLA-DR molecules to bind diverse peptide repertoires. , 1997, Journal of immunology.

[14]  Toshinori Munakata,et al.  Rule extraction from expert heuristics: A comparative study of rough sets with neural networks and ID3 , 2002, Eur. J. Oper. Res..

[15]  Li Xiangyang,et al.  A novel self-optimizing approach for knowledge acquisition , 2000 .