Machine learning competition in immunology – Prediction of HLA class I binding peptides

Guang Lan Zhang, Hifzur Rahman Ansari, Phil Bradley, Gavin C. Cawley, Tomer Hertz, Xihao Hu, Nebojsa Jojic, Yohan Kim, Oliver Kohlbacher, Ole Lund, Claus Lundegaard, Craig A. Magaret, Morten Nielsen, Harris Papadopoulos, G.P.S. Raghava, Vider-Shalit Tal, Li C. Xue, Chen Yanover, Shanfeng Zhu, Michael T. Rock & 5 others James E. Crowe, Christos Panayiotou, Marios M. Polycarpou, Włodzisław Duch, Vladimir Brusic

    Research output: Contribution to journalJournal articleResearchpeer-review

    Abstract

    Experimental studies of immune system and related applications such as characterization of immune responses against pathogens, vaccine design, or optimization of therapies are combinatorially complex, time-consuming and expensive. The main methods for large-scale identification of T-cell epitopes from pathogens or cancer proteomes involve either reverse immunology or high-throughput mass spectrometry (HTMS). Reverse immunology approaches involve pre-screening of proteomes by computational algorithms, followed by experimental validation of selected targets ( [Mora et al., 2006], [De Groot et al., 2008] and [Larsen et al., 2010]). HTMS involves HLA typing, immunoaffinity chromatography of HLA molecules, HLA extraction, and chromatography combined with tandem mass spectrometry, followed by the application of computational algorithms for peptide characterization (Bassani-Sternberg et al., 2010). Hundreds of naturally processed HLA class I associated peptides have been identified in individual studies using HTMS in normal (Escobar et al., 2008), cancer ( [Antwi et al., 2009] and [Bassani-Sternberg et al., 2010]), autoimmunity-related (Ben Dror et al., 2010), and infected samples (Wahl et al, 2010). Computational algorithms are essential steps in high-throughput identification of T-cell epitope candidates using both reverse immunology and HTMS approaches. Peptide binding to MHC molecules is the single most selective step in defining T cell epitope and the accuracy of computational algorithms for prediction of peptide binding, therefore, determines the accuracy of the overall method. Computational predictions of peptide binding to HLA, both class I and class II, use a variety of algorithms ranging from binding motifs to advanced machine learning techniques ( [Brusic et al., 2004] and [Lafuente and Reche, 2009]) and standards for their assessments have been developed. The assessments of computational servers that predict peptide binding to several common HLA class I alleles have been performed by different groups (see [Peters et al., 2006], [Lin et al., 2008] and [Gowthaman et al., 2010]). Some of these models were reported to be highly accurate while others need improvement.
    Original languageEnglish
    JournalJournal of Immunological Methods
    Volume374
    Issue number1-2
    Pages (from-to)1-4
    ISSN0022-1759
    DOIs
    Publication statusPublished - 2011

    Keywords

    • Computational models
    • Peptide binding
    • Human leukocyte antigen
    • HLA
    • Predictions
    • Machine learning

    Cite this

    Zhang, G. L., Ansari, H. R., Bradley, P., Cawley, G. C., Hertz, T., Hu, X., ... Brusic, V. (2011). Machine learning competition in immunology – Prediction of HLA class I binding peptides. Journal of Immunological Methods, 374(1-2), 1-4. https://doi.org/10.1016/j.jim.2011.09.010
    Zhang, Guang Lan ; Ansari, Hifzur Rahman ; Bradley, Phil ; Cawley, Gavin C. ; Hertz, Tomer ; Hu, Xihao ; Jojic, Nebojsa ; Kim, Yohan ; Kohlbacher, Oliver ; Lund, Ole ; Lundegaard, Claus ; Magaret, Craig A. ; Nielsen, Morten ; Papadopoulos, Harris ; Raghava, G.P.S. ; Tal, Vider-Shalit ; Xue, Li C. ; Yanover, Chen ; Zhu, Shanfeng ; Rock, Michael T. ; Crowe, James E. ; Panayiotou, Christos ; Polycarpou, Marios M. ; Duch, Włodzisław ; Brusic, Vladimir. / Machine learning competition in immunology – Prediction of HLA class I binding peptides. In: Journal of Immunological Methods. 2011 ; Vol. 374, No. 1-2. pp. 1-4.
    @article{506be7f6badb4515a3fc1536cde353f8,
    title = "Machine learning competition in immunology – Prediction of HLA class I binding peptides",
    abstract = "Experimental studies of immune system and related applications such as characterization of immune responses against pathogens, vaccine design, or optimization of therapies are combinatorially complex, time-consuming and expensive. The main methods for large-scale identification of T-cell epitopes from pathogens or cancer proteomes involve either reverse immunology or high-throughput mass spectrometry (HTMS). Reverse immunology approaches involve pre-screening of proteomes by computational algorithms, followed by experimental validation of selected targets ( [Mora et al., 2006], [De Groot et al., 2008] and [Larsen et al., 2010]). HTMS involves HLA typing, immunoaffinity chromatography of HLA molecules, HLA extraction, and chromatography combined with tandem mass spectrometry, followed by the application of computational algorithms for peptide characterization (Bassani-Sternberg et al., 2010). Hundreds of naturally processed HLA class I associated peptides have been identified in individual studies using HTMS in normal (Escobar et al., 2008), cancer ( [Antwi et al., 2009] and [Bassani-Sternberg et al., 2010]), autoimmunity-related (Ben Dror et al., 2010), and infected samples (Wahl et al, 2010). Computational algorithms are essential steps in high-throughput identification of T-cell epitope candidates using both reverse immunology and HTMS approaches. Peptide binding to MHC molecules is the single most selective step in defining T cell epitope and the accuracy of computational algorithms for prediction of peptide binding, therefore, determines the accuracy of the overall method. Computational predictions of peptide binding to HLA, both class I and class II, use a variety of algorithms ranging from binding motifs to advanced machine learning techniques ( [Brusic et al., 2004] and [Lafuente and Reche, 2009]) and standards for their assessments have been developed. The assessments of computational servers that predict peptide binding to several common HLA class I alleles have been performed by different groups (see [Peters et al., 2006], [Lin et al., 2008] and [Gowthaman et al., 2010]). Some of these models were reported to be highly accurate while others need improvement.",
    keywords = "Computational models, Peptide binding, Human leukocyte antigen, HLA, Predictions, Machine learning",
    author = "Zhang, {Guang Lan} and Ansari, {Hifzur Rahman} and Phil Bradley and Cawley, {Gavin C.} and Tomer Hertz and Xihao Hu and Nebojsa Jojic and Yohan Kim and Oliver Kohlbacher and Ole Lund and Claus Lundegaard and Magaret, {Craig A.} and Morten Nielsen and Harris Papadopoulos and G.P.S. Raghava and Vider-Shalit Tal and Xue, {Li C.} and Chen Yanover and Shanfeng Zhu and Rock, {Michael T.} and Crowe, {James E.} and Christos Panayiotou and Polycarpou, {Marios M.} and Włodzisław Duch and Vladimir Brusic",
    year = "2011",
    doi = "10.1016/j.jim.2011.09.010",
    language = "English",
    volume = "374",
    pages = "1--4",
    journal = "Journal of Immunological Methods",
    issn = "0022-1759",
    publisher = "Elsevier",
    number = "1-2",

    }

    Zhang, GL, Ansari, HR, Bradley, P, Cawley, GC, Hertz, T, Hu, X, Jojic, N, Kim, Y, Kohlbacher, O, Lund, O, Lundegaard, C, Magaret, CA, Nielsen, M, Papadopoulos, H, Raghava, GPS, Tal, V-S, Xue, LC, Yanover, C, Zhu, S, Rock, MT, Crowe, JE, Panayiotou, C, Polycarpou, MM, Duch, W & Brusic, V 2011, 'Machine learning competition in immunology – Prediction of HLA class I binding peptides', Journal of Immunological Methods, vol. 374, no. 1-2, pp. 1-4. https://doi.org/10.1016/j.jim.2011.09.010

    Machine learning competition in immunology – Prediction of HLA class I binding peptides. / Zhang, Guang Lan; Ansari, Hifzur Rahman; Bradley, Phil; Cawley, Gavin C.; Hertz, Tomer; Hu, Xihao; Jojic, Nebojsa; Kim, Yohan; Kohlbacher, Oliver; Lund, Ole; Lundegaard, Claus; Magaret, Craig A.; Nielsen, Morten; Papadopoulos, Harris; Raghava, G.P.S.; Tal, Vider-Shalit; Xue, Li C.; Yanover, Chen; Zhu, Shanfeng; Rock, Michael T.; Crowe, James E.; Panayiotou, Christos; Polycarpou, Marios M.; Duch, Włodzisław; Brusic, Vladimir.

    In: Journal of Immunological Methods, Vol. 374, No. 1-2, 2011, p. 1-4.

    Research output: Contribution to journalJournal articleResearchpeer-review

    TY - JOUR

    T1 - Machine learning competition in immunology – Prediction of HLA class I binding peptides

    AU - Zhang, Guang Lan

    AU - Ansari, Hifzur Rahman

    AU - Bradley, Phil

    AU - Cawley, Gavin C.

    AU - Hertz, Tomer

    AU - Hu, Xihao

    AU - Jojic, Nebojsa

    AU - Kim, Yohan

    AU - Kohlbacher, Oliver

    AU - Lund, Ole

    AU - Lundegaard, Claus

    AU - Magaret, Craig A.

    AU - Nielsen, Morten

    AU - Papadopoulos, Harris

    AU - Raghava, G.P.S.

    AU - Tal, Vider-Shalit

    AU - Xue, Li C.

    AU - Yanover, Chen

    AU - Zhu, Shanfeng

    AU - Rock, Michael T.

    AU - Crowe, James E.

    AU - Panayiotou, Christos

    AU - Polycarpou, Marios M.

    AU - Duch, Włodzisław

    AU - Brusic, Vladimir

    PY - 2011

    Y1 - 2011

    N2 - Experimental studies of immune system and related applications such as characterization of immune responses against pathogens, vaccine design, or optimization of therapies are combinatorially complex, time-consuming and expensive. The main methods for large-scale identification of T-cell epitopes from pathogens or cancer proteomes involve either reverse immunology or high-throughput mass spectrometry (HTMS). Reverse immunology approaches involve pre-screening of proteomes by computational algorithms, followed by experimental validation of selected targets ( [Mora et al., 2006], [De Groot et al., 2008] and [Larsen et al., 2010]). HTMS involves HLA typing, immunoaffinity chromatography of HLA molecules, HLA extraction, and chromatography combined with tandem mass spectrometry, followed by the application of computational algorithms for peptide characterization (Bassani-Sternberg et al., 2010). Hundreds of naturally processed HLA class I associated peptides have been identified in individual studies using HTMS in normal (Escobar et al., 2008), cancer ( [Antwi et al., 2009] and [Bassani-Sternberg et al., 2010]), autoimmunity-related (Ben Dror et al., 2010), and infected samples (Wahl et al, 2010). Computational algorithms are essential steps in high-throughput identification of T-cell epitope candidates using both reverse immunology and HTMS approaches. Peptide binding to MHC molecules is the single most selective step in defining T cell epitope and the accuracy of computational algorithms for prediction of peptide binding, therefore, determines the accuracy of the overall method. Computational predictions of peptide binding to HLA, both class I and class II, use a variety of algorithms ranging from binding motifs to advanced machine learning techniques ( [Brusic et al., 2004] and [Lafuente and Reche, 2009]) and standards for their assessments have been developed. The assessments of computational servers that predict peptide binding to several common HLA class I alleles have been performed by different groups (see [Peters et al., 2006], [Lin et al., 2008] and [Gowthaman et al., 2010]). Some of these models were reported to be highly accurate while others need improvement.

    AB - Experimental studies of immune system and related applications such as characterization of immune responses against pathogens, vaccine design, or optimization of therapies are combinatorially complex, time-consuming and expensive. The main methods for large-scale identification of T-cell epitopes from pathogens or cancer proteomes involve either reverse immunology or high-throughput mass spectrometry (HTMS). Reverse immunology approaches involve pre-screening of proteomes by computational algorithms, followed by experimental validation of selected targets ( [Mora et al., 2006], [De Groot et al., 2008] and [Larsen et al., 2010]). HTMS involves HLA typing, immunoaffinity chromatography of HLA molecules, HLA extraction, and chromatography combined with tandem mass spectrometry, followed by the application of computational algorithms for peptide characterization (Bassani-Sternberg et al., 2010). Hundreds of naturally processed HLA class I associated peptides have been identified in individual studies using HTMS in normal (Escobar et al., 2008), cancer ( [Antwi et al., 2009] and [Bassani-Sternberg et al., 2010]), autoimmunity-related (Ben Dror et al., 2010), and infected samples (Wahl et al, 2010). Computational algorithms are essential steps in high-throughput identification of T-cell epitope candidates using both reverse immunology and HTMS approaches. Peptide binding to MHC molecules is the single most selective step in defining T cell epitope and the accuracy of computational algorithms for prediction of peptide binding, therefore, determines the accuracy of the overall method. Computational predictions of peptide binding to HLA, both class I and class II, use a variety of algorithms ranging from binding motifs to advanced machine learning techniques ( [Brusic et al., 2004] and [Lafuente and Reche, 2009]) and standards for their assessments have been developed. The assessments of computational servers that predict peptide binding to several common HLA class I alleles have been performed by different groups (see [Peters et al., 2006], [Lin et al., 2008] and [Gowthaman et al., 2010]). Some of these models were reported to be highly accurate while others need improvement.

    KW - Computational models

    KW - Peptide binding

    KW - Human leukocyte antigen

    KW - HLA

    KW - Predictions

    KW - Machine learning

    U2 - 10.1016/j.jim.2011.09.010

    DO - 10.1016/j.jim.2011.09.010

    M3 - Journal article

    VL - 374

    SP - 1

    EP - 4

    JO - Journal of Immunological Methods

    JF - Journal of Immunological Methods

    SN - 0022-1759

    IS - 1-2

    ER -