Vu Anh Tran

PAAQD: Predicting immunogenicity of MHC class I binding peptides using amino acid pairwise contact potentials and quantum topological molecular similarity descriptors.

Prediction of peptide immunogenicity is a promising approach for novel vaccine discovery. Conventionally, epitope prediction methods have been developed to accelerate the process of vaccine production by searching for candidate peptides from pathogenic proteins. However, recent studies revealed that peptides with high binding affinity to major histocompatibility complex molecules (MHCs) do not always result in high immunogenicity. Therefore, it is promising to predict the peptide immunogenicity rather than epitopes in order to discover new vaccines more effectively. To this end, we developed a novel T-cell reactivity predictor which we call PAAQD. Nonapeptides were encoded numerically, using combining information of amino acid pairwise contact potentials (AAPPs) and quantum topological molecular similarity (QTMS) descriptors. Encoded data were used in the construction of our classification model. Our numerical experiments suggested that the predictive performance of PAAQD is at least comparable with POPISK, one of the pioneering techniques for T-cell reactivity prediction. Also, our experiment suggested that the first and eighth positions of nonapeptides are the most important for immunogenicity and most of the anchor residues in epitope prediction were not important in T-cell reactivity prediction. The R implementation of PAAQD is available at http://pirun.ku.ac.th/~fsciiok/PAAQD.rar.

EpicCapo: epitope prediction using combined information of amino acid pairwise contact potentials and HLA-peptide contact site information

BackgroundEpitope identification is an essential step toward synthetic vaccine development since epitopes play an important role in activating immune response. Classical experimental approaches are laborious and time-consuming, and therefore computational methods for generating epitope candidates have been actively studied. Most of these methods, however, are based on sophisticated nonlinear techniques for achieving higher predictive performance. The use of these techniques tend to diminish their interpretability with respect to binding potential: that is, they do not provide much insight into binding mechanisms.ResultsWe have developed a novel epitope prediction method named EpicCapo and its variants, EpicCapo+ and EpicCapo+REF. Nonapeptides were encoded numerically using a novel peptide-encoding scheme for machine learning algorithms by utilizing 40 amino acid pairwise contact potentials (referred to as AAPPs throughout this paper). The predictive performances of EpicCapo+ and EpicCapo+REF outperformed other state-of-the-art methods without losing interpretability. Interestingly, the most informative AAPPs estimated by our study were those developed by Micheletti and Simons while previous studies utilized two AAPPs developed by Miyazawa & Jernigan and Betancourt & Thirumalai. In addition, we found that all amino acid positions in nonapeptides could effect on performances of the predictive models including non-anchor positions. Finally, EpicCapo+REF was applied to identify candidates of promiscuous epitopes. As a result, 67.1% of the predicted nonapeptides epitopes were consistent with preceding studies based on immunological experiments.ConclusionsOur method achieved high performance in testing with benchmark datasets. In addition, our study identified a number of candidates of promiscuous CTL epitopes consistent with previously reported immunological experiments. We speculate that our techniques may be useful in the development of new vaccines. The R implementation of EpicCapo+REF is available athttp://pirun.ku.ac.th/~fsciiok/EpicCapoREF.zip. Datasets are available athttp://pirun.ku.ac.th/~fsciiok/Datasets.zip.

Inference of domain–domain interactions by matrix factorisation and domain–level features

In the development of new drugs and improved treatment of diseases, it is essential to understand molecular networks in living organism. Especially, it is important to identify interacting domains among proteins to elucidate hidden functions for protein–protein interactions (PPIs). To date, a number of computational methods have been developed for predicting domain–domain interactions (DDIs) from known PPIs. However, they often contain a large number of false positives while the number of known structures of protein complexes is limited. In this study, we aim to develop a new method of predicting DDIs by a link prediction approach. By using a learning model including low rank matrices as latent features in combination with biological features and topological features of the domain network, the experimental results showed that our method achieved a good performance and the predicted DDIs have high fraction sharing rate with the ones known as true in gold–standard databases.