Journal of Biological Physics

, Volume 28, Issue 2, pp 183–194 | Cite as

Prediction of MHC Class I Binding Peptides by a Query Learning Algorithm Based on Hidden Markov Models

  • Keiko Udaka
  • Hiroshi Mamitsuka
  • Yukinobu Nakaseko
  • Naoki Abe


A query learning algorithm based on hidden Markov models (HMMs) isdeveloped to design experiments for string analysis and prediction of MHCclass I binding peptides. Query learning is introduced to aim at reducingthe number of peptide binding data for training of HMMs. A multiple numberof HMMs, which will collectively serve as a committee, are trained withbinding data and used for prediction in real-number values. The universeof peptides is randomly sampled and subjected to judgement by the HMMs.Peptides whose prediction is least consistent among committee HMMs aretested by experiment. By iterating the feedback cycle of computationalanalysis and experiment the most wanted information is effectivelyextracted. After 7 rounds of active learning with 181 peptides in all,predictive performance of the algorithm surpassed the so far bestperforming matrix based prediction. Moreover, by combining the bothmethods binder peptides (log Kd < -6) could be predicted with84% accuracy. Parameter distribution of the HMMs that can be inspectedvisually after training further offers a glimpse of dynamic specificity ofthe MHC molecules.

algorithm binding experimental design MHC class I molecules peptides prediction query learning specificity string analysis 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Falk, K., Roetzschke, O., Stevanovic, S., Jung, G. and Rammensee, H.: Allele-specific motifs revealed by sequencing of self-peptides eluted from MHC molecules, Nature 351 (1991), 290–296.Google Scholar
  2. 2.
    Rammensee, H., Friede, T. and Stevanovic, S.: MHC ligands and peptide motifs: first listing, Immunogenetics 41 (1995), 178–228.PubMedGoogle Scholar
  3. 3.
    Ruppert, J., et al.: Prominent role of secondary anchor residues in peptide binding to HLA-A2.1 molecules, Cell 74 (1993), 929–937.Google Scholar
  4. 4.
    Parker, K., Bednarek, M. and Je, C.: Scheme for ranking potential HLA-A2 binding peptides based on independent binding of individual peptide side-chains, J. Immunol. 152 (1994), 163–175.Google Scholar
  5. 5.
    Hammer, J., et al.: Precise prediction of MHC class II-peptide interaction based on peptide side chain scanning, J. Exp. Med. 180 (1994), 2353–2358.Google Scholar
  6. 6.
    Stryhn, A., et al.: Peptide binding specificity of major histocompatibility complex class I resolved into an array of apparently independent subspecificities: quantitation by peptide libraries and improved prediction of binding, Eur. J. Immunol. 26 (1996), 1911–1918.Google Scholar
  7. 7.
    Udaka, K., Wiesmuller, K.-H., Kienle, S., Jung, G. and Walden, P.: Tolerance to amino acid variations in peptides binding to the MHC class I protein H-2Kb, J. Biol. Chem. 270 (1995), 24130–24134.Google Scholar
  8. 8.
    Udaka, K., Wiesmueller, K.-H., Kienle, S., Jung, G. and Walden, P.: Decrypting the structure of MHC-I restricted CTL epitopes with complex peptide libraries, J. Exp. Med. 181 (1995), 2097–2108.PubMedGoogle Scholar
  9. 9.
    Brusic, V., Schoenbach, C., Takiguchi, M., Ciesielski, V. and Harrison, L.: Application of genetic search in derivation of matrix models of peptide binding to MHC molecules, ISMB (1997), 75–83.Google Scholar
  10. 10.
    Gulukota, K., Sidney, J., Sette, A. and DeLisi, C.: Two complementary methods for predicting peptides binding major histocompatibility complex molecules, J. Mol. Biol. 267 (1997), 1258–1267.Google Scholar
  11. 11.
    Honeyman, M., Brusic, V., Stone, N. and Harrison, L.: Neural network-based prediction of candidate T-cell epitopes, Nat. Biotechnol. 16 (1998), 966–969.Google Scholar
  12. 12.
    Rammensee, H., Bachmann, J., Emmerich, N., Bachor, 0. and Stevanovic, S.: SYFPEITHI: Database for MHC ligands and peptide motifs, Immunogenetics 50 (1999), 213–219.Google Scholar
  13. 13.
    Andersen, M., et al.: Poor correspondence between predicted and experimental binding of peptides to class I MHC molecules, Tissue Antigens 55 (2000), 519–531.Google Scholar
  14. 14.
    Udaka, K., et al.: An automated prediction of MHC class I-binding peptides based on positional scanning with peptide libraries, Immunogenetics 51 (2000), 816–828.Google Scholar
  15. 15.
    Madden, D., Garboczi, D. and Wiley, D.: The antigenic identity of peptide-MHC complexes: A comparison of the conformations of five viral peptides by HLA-A2, Cell 75 (1993), 693–708.Google Scholar
  16. 16.
    Mamitsuka, H.: Predicting peptides that bind to MHC molecules using supervised learning of Hidden Markov Models, Proteins 33 (1998), 460–474.Google Scholar
  17. 17.
    Brusic, V., Rudy, G., Honeyman, M., Hammer, J. and Harrison, L.: Prediction of MHC class IIbinding peptides using an evolutionary algorithm and artificial neural network, Bioinformatics 14 (1998), 121–130.Google Scholar
  18. 18.
    Brusic, V., Rudy, G., Kyne, A. and Harrison, L.: MHCPEP – A database of MHC-binding peptides: update 1995, Nucl. Acids Res. 24 (1996), 242–244.Google Scholar
  19. 19.
    Abe, N. and Mamitsuka, H.: In the fifteenth international conference on machine learning, Morgan Kaufmann, Madison, Wisconsin, 1998.Google Scholar
  20. 20.
    Mamitsuka, H. and Abe, N.: In the seventeenth international conference on machine learning, Morgan Kaufmann, Stanford, Ca., 2000.Google Scholar
  21. 21.
    Baldi, P., Chauvin, Y., Hunkapiler, T. and McClure, M.A.: HiddenMarkov models of biological primary sequence information, Proc. Natl. Acad. Sci. USA 91 (1994), 1059–1063.Google Scholar
  22. 22.
    Krogh, A., Brown, M., Mian, I.S., Sjolander, K. and Haussler, D.: Hidden Markov models in computational biology: applications to protein modeling, J. Mol. Biol. 235 (1994), 1501–1531.Google Scholar
  23. 23.
    Park, J., et al.: Sequence comparisons using multiple sequences detect twice as many remote homologues as pairwise methods, J. Mol. Biol. 284 (1998), 1201–1210. 194 K. UDAKA ET AL.Google Scholar
  24. 24.
    Krogh, A., I.S., M. and D., H.: A hidden Markov model that finds genes in E. coli DNA, Nucl. Acids Res. 22 (1994), 4768–4778.Google Scholar
  25. 25.
    Kearns, M. and Vazirani, U.: An introduction to computational learning theory, MIT Press, Cambridge, Ma., 1994.Google Scholar
  26. 26.
    Mamitsuka, H.: A learning method of hidden Markov Models for sequence discrimination, J. Comput. Biol. 3 (1996), 361–363.Google Scholar

Copyright information

© Kluwer Academic Publishers 2002

Authors and Affiliations

  • Keiko Udaka
    • 1
  • Hiroshi Mamitsuka
    • 2
  • Yukinobu Nakaseko
    • 1
  • Naoki Abe
    • 2
  1. 1.Department of BiophysicsKyoto UniversityJapan
  2. 2.Theory NEC Laboratory, RWCP (Real Worid Computing Partnership), c/o Internet Systems Research LaboratoriesNEC corporationJapan

Personalised recommendations