Use of BONSAI decision trees for the identification of potential MHC Class I peptide epitope motifs

Savoie, Christopher J.; Kamikawaji, Nobuhiro; Sasazuki, Takehiko; Kuhara, Satoru

doi:10.1142/9789814447300_0018

Cited by 24 publications

(16 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In this approach, the DT is used to construct a graph model of the MHC-peptide binding motif, which subsequently can be used to decide whether a test peptide fits into that motif. Savoie et al [56] were the first to report the use of BONSAI DTs to predict peptide binding for HLA-A*0201. A similar tree-structured technique was later reported by Segal et al [57] to predict peptide binding to K b (a mouse MHCI molecule).…”

Section: Machine Learning-based Motifsmentioning

confidence: 99%

Prediction of MHC-Peptide Binding: A Systematic and Comprehensive Overview

Lafuente¹,

Reche

2009

CPD

136

View full text Add to dashboard Cite

T cell immune responses are driven by the recognition of peptide antigens (T cell epitopes) that are bound to major histocompatibility complex (MHC) molecules. T cell epitope immunogenicity is thus contingent on several events, including appropriate and effective processing of the peptide from its protein source, stable peptide binding to the MHC molecule, and recognition of the MHC-bound peptide by the T cell receptor. Of these three hallmarks, MHC-peptide binding is the most selective event that determines T cell epitopes. Therefore, prediction of MHC-peptide binding constitutes the principal basis for anticipating potential T cell epitopes. The tremendous relevance of epitope identification in vaccine design and in the monitoring of T cell responses has spurred the development of many computational methods for predicting MHC-peptide binding that improve the efficiency and economics of T cell epitope identification. In this report, we will systematically examine the available methods for predicting MHC-peptide binding and discuss their most relevant advantages and drawbacks.

show abstract

Section: Machine Learning-based Motifsmentioning

confidence: 99%

Prediction of MHC-Peptide Binding: A Systematic and Comprehensive Overview

Lafuente¹,

Reche

2009

CPD

136

View full text Add to dashboard Cite

show abstract

“…The general methods, such as classification and regression trees, Neural Networks (NN), Hidden Markov Models (HMM), or Support Vector Machines (SVM), multivariate statistical approaches and decision trees, take into account possible combined influences of multiple residues in different positions in a peptide. Decision trees [11] were used to predict MHC-peptide binding. Decision trees use a natural and intuitive way to classify a pattern through a sequence of questions in which the next question asked depends on the answer to the current question [12].…”

Section: Review Of Current Machine Learning Methods In Predicting Hla-mentioning

confidence: 99%

Extreme Learning Machine for Predicting HLA-Peptide Binding

Handoko

Keong

Soon

et al. 2006

Advances in Neural Networks - ISNN 2006

View full text Add to dashboard Cite

Abstract. Machine learning techniques have been recognized as powerful tools for learning from data. One of the most popular learning techniques, the BackPropagation (BP) Artificial Neural Networks, can be used as a computer model to predict peptides binding to the Human Leukocyte Antigens (HLA). The major advantage of computational screening is that it reduces the number of wet-lab experiments that need to be performed, significantly reducing the cost and time. A recently developed method, Extreme Learning Machine (ELM), which has superior properties over BP has been investigated to accomplish such tasks. In our work, we found that the ELM is as good as, if not better than, the BP in term of time complexity, accuracy deviations across experiments, andmost importantly -prevention from over-fitting for prediction of peptide binding to HLA.

show abstract

“…Data-driven approaches include statistical methods based on experimental peptide binding measurements. These methods include binding motifs (Rammensee et al, 1993), quantitative matrices (Parker et al, 1994;Singh and Raghava, 2003;Reche and Reinherz, 2005;Peters and Sette, 2005), artificial neural networks (ANN) (Honeyman et al, 1998;Christensen et al, 2003), hidden Markov models (HMM) (Mamitsuka, 1998;Brusic et al, 2002), decision trees (Savoie et al, 1999;Segal et al, 2001), discriminant analysis (Mallios, 2001), multivariate regression (Lin et al, 2004), ensemble classifier (Xiao and Segal, 2005), support vector machines (SVM) (Donnes and Elofsson, 2002;Zhao et al, 2003;Bhasin and Raghava, 2004;Riedesel et al, 2004;Bozic et al, 2005;Liu et al, 2006;Cui et al, 2007), and biosupport vector machine which is modified from a conventional support vector machine by introducing a biobasis function so that the non-numerical attributes of amino acids can be recognized without a feature extraction process (Yang and Johnson, 2005). Recently a structure-and sequence-based method was reported, in which residue-based energy terms from the molecular dynamics simulations are used as features to train SVM prediction models for peptide/MHC class I binding (Antes et al, 2006).…”

Section: Introductionmentioning

confidence: 99%

Prediction of supertype-specific HLA class I binding peptides using support vector machines

Zhang

Božić

Kwoh

et al. 2007

Journal of Immunological Methods

View full text Add to dashboard Cite

Experimental approaches for identifying T-cell epitopes are time-consuming, costly and not applicable to the large scale screening. Computer modeling methods can help to minimize the number of experiments required, enable a systematic scanning for candidate major histocompatibility complex (MHC) binding peptides and thus speed up vaccine development. We developed a prediction system based on a novel data representation of peptide/MHC interaction and support vector machines (SVM) for prediction of peptides that promiscuously bind to multiple Human Leukocyte Antigen (HLA, human MHC) alleles belonging to a HLA supertype. Ten-fold cross-validation results showed that the overall performance of SVM models is improved in comparison to our previously published methods based on hidden Markov models (HMM) and artificial neural networks (ANN), also confirmed by blind testing. At specificity 0.90, sensitivity values of SVM models were 0.90 and 0.92 for HLA-A2 and -A3 dataset respectively. Average area under the receiver operating curve (A ROC ) of SVM models in blind testing are 0.89 and 0.92 for HLA-A2 and -A3 datasets. A ROC of HLA-A2 and -A3 SVM models were 0.94 and 0.95, validated using a full overlapping study of 9-mer peptides from human papillomavirus type 16 E6 and E7 proteins. In addition, a large-scale experimental dataset has been used to validate HLA-A2 and -A3 SVM models. The SVM prediction models were integrated into a web-based computational system MULTIPRED1, accessible at antigen.i2r.a-star.edu.sg/multipred1/.

show abstract

Use of BONSAI decision trees for the identification of potential MHC Class I peptide epitope motifs

Cited by 24 publications

References 12 publications

Prediction of MHC-Peptide Binding: A Systematic and Comprehensive Overview

Prediction of MHC-Peptide Binding: A Systematic and Comprehensive Overview

Extreme Learning Machine for Predicting HLA-Peptide Binding

Prediction of supertype-specific HLA class I binding peptides using support vector machines

Contact Info

Product

Resources

About