Modeling and prediction of binding affinities between the human amphiphysin SH3 domain and its peptide ligands using genetic algorithm-Gaussian processes

In this article, we discuss the application of the Gaussian process (GP) and other statistical methods (PLS, ANN, and SVM) for the modeling and prediction of binding affinities between the human amphiphysin SH3 domain and its peptide ligands. Divided physicochemical property scores of amino acids, involving significant hydrogen bond, electronic, hydrophobic, and steric properties, was used to characterize the peptide structures, and quantitative structure-affinity relationship models were then constructed by PLS, ANN, SVM, and GP coupled with genetic algorithm-variable selection. The results show that: (i) since the significant flexibility and high complexity possessed in polypeptide structures, linear PLS method was incapable of fulfilling a satisfying behavior on SH3 domain binding peptide dataset; (ii) the overfitting involved in training process has decreased the predictive power of ANN model to some extent; (iii) both SVM and GP have a good performance for SH3 domain binding peptide dataset. Moreover, by combining linear and nonlinear terms in the covariance function, the GP is capable of handling linear and nonlinear-hybrid relationship, and which thus obtained a more stable and predictable model than SVM. Analyses of GP models showed that diversified properties contribute remarkable effect to the interactions between the SH3 domain and the peptides. Particularly, steric property and hydrophobicity of P(2), electronic property of P(0), and electronic property and hydrogen bond property of P(-3) in decapeptide (P(4)P(3)P(2)P(1)P(0)P(-1)P(-2)P(-3)P(-4)P(-5)) significantly contribute to the binding affinities of SH3 domain-peptide interactions.     

Gaussian process: a promising approach for the modeling and prediction of Peptide binding affinity to MHC proteins

On the basis of Bayesian probabilistic inference, Gaussian process (GP) is a powerful machine learning method for nonlinear classification and regression, but has only very limited applications in the new areas of computational vaccinology and immunoinformatics. In the current work, we present a paradigmatic study of using GP regression technique to quantitatively model and predict the binding affinities of over 7000 immunodominant peptide epitopes to six types of human major histocompatibility complex (MHC) proteins. In this procedure, the sequence patterns of diverse peptides are characterized quantitatively and the resulting variables are then correlated with the experimentally measured affinities between different MHC and their peptide ligands, by using a linearity- and nonlinearity-hybrid GP approach. We also make systematical comparisons between the GP and two sophisticated modeling methods as partial least square (PLS) regression and support vector machine (SVM) with respect to their fitting ability, predictive power and generalization capability. The results suggest that GP could be a new and effective tool for the modeling and prediction of MHC-peptide interactions and would be promising in the field of computer-aided vaccine design (CAVD).     

Connecting membrane fluidity and surface charge to pore-forming antimicrobial peptides resistance by an ANN-based predictive model

Efficiency of antibacterial chemotherapy is gradually more challenged by the emergence of pathogenic strains exhibiting high levels of antibiotic resistance. Pore-forming antimicrobial peptides (PF-AMPs) such as alamethicin (Alm) are therefore in the focus of extensive research efforts. In the present study, an artificial neural network (ANN)-based quantitative structure–activity relationship (SAR) modeling of membrane phospholipids vs. PF-AMPs, in context to membrane fluidity and surface charge, was carried out. We observed that the potency of PF-AMPs depends on the fatty acyl chain and polar head group of phospholipids. Alm showed surface interactions with zwitterionic phospholipids however could penetrate deeper inside the hydrophobic core of anionic membranes. Here, the resistance developed in bacterial cells was coupled to membrane fluidity and surface charge, and simultaneously, these principles could be applied for combating resistance against PF-AMPs. The correlation coefficient between observed CR and predicted CR using ANN was found to be 0.757. Thus, ANN could be used as a reliable modeling method for predicting CR, given the structure of the biomimetic membrane in terms of membrane fluidity and surface charge. Fully explored mechanisms of resistance, a forward modeling step in the design cycle of AMPs, can be cross-linked to the inward modeling using ANN to complete the peptide design cycle. The SAR between membrane phospholipids and PF-AMPs could furnish valuable information regarding their design to provide us efficacious peptides against premier pathogens. So far, this is the only report available to predict and quantify interactions of PF-AMPs with membrane phospholipids.     

A neural network model approach to the study of human TAP transporter

We used an artificial neural network (ANN) computer model to study peptide binding to the human transporter associated with antigen processing (TAP). After validation, an ANN model of TAP-peptide binding was used to mine a database of HLA-binding peptides to elucidate patterns of TAP binding. The affinity of HLA-binding peptides for TAP was found to differ according to the HLA supertype concerned: HLA-B27, -A3 or -A24 binding peptides had high, whereas HLA-A2, -B7 or -B8 binding peptides had low affinity for TAP. These results support the idea that TAP and particular HLA molecules may have co-evolved for efficient peptide processing and presentation. The strong similarity between the sets of peptides bound by TAP or HLA-B27 suggests functional co-evolution whereas the lack of a relationship between the sets of peptides bound by TAP or HLA-A2 is against these particular molecules having co-evolved. In support of these conclusions, the affinities of HLA-A2 and HLA-B7 binding peptides for TAP show similar distributions to that of randomly generated peptides. On the basis of these results we propose that HLA alleles constitute two separate classes: those that are TAP-efficient for peptide loading (HLA-B27, -A3 and -A24) and those that are TAP-inefficient (HLA-A2, -B7 and -B8). Computer modelling can be used to complement laboratory experiments and thereby speed up knowledge discovery in biology. In particular, we provide evidence that large-scale experiments can be avoided by combining initial experimental data with limited laboratory experiments sufficient to develop and validate appropriate computer models. These models can then be used to perform large-scale simulated experiments the results of which can then be validated by further small-scale laboratory experiments.     

Using multidimensional patterns of amino acid attributes for QSAR analysis of peptides

On the basis of exploratory factor analysis, six multidimensional patterns of 516 amino acid attributes, namely, factor analysis scales of generalized amino acid information (FASGAI) involving hydrophobicity, alpha and turn propensities, bulky properties, compositional characteristics, local flexibility and electronic properties, are proposed to represent structures of 48 bitter-tasting dipeptides and 58 angiotensin-converting enzyme inhibitors. Characteristic parameters related to bioactivities of the peptides studied are selected by genetic algorithm, and quantitative structure–activity relationship (QSAR) models are constructed by partial least square (PLS). Our results by a leave-one-out cross validation are compared with the previously known structure representation method and are shown to give slightly superior or comparative performance. Further, two data sets are divided into training sets and test sets to validate the characterization repertoire of FASGAI. Performance of the PLS models developed by training samples by a leave-one-out cross validation and external validation for test samples are satisfying. These results demonstrate that FASGAI is an effective representation technique of peptide structures, and that FASGAI vectors have many preponderant characteristics such as straightforward physicochemical information, high characterization competence and easy manipulation. They can be further applied to investigate the relationship between structures and functions of various peptides, even proteins.     

POPI: predicting immunogenicity of MHC class I binding peptides by mining informative physicochemical properties

MOTIVATION: Both modeling of antigen-processing pathway including major histocompatibility complex (MHC) binding and immunogenicity prediction of those MHC-binding peptides are essential to develop a computer-aided system of peptide-based vaccine design that is one goal of immunoinformatics. Numerous studies have dealt with modeling the immunogenic pathway but not the intractable problem of immunogenicity prediction due to complex effects of many intrinsic and extrinsic factors. Moderate affinity of the MHC-peptide complex is essential to induce immune responses, but the relationship between the affinity and peptide immunogenicity is too weak to use for predicting immunogenicity. This study focuses on mining informative physicochemical properties from known experimental immunogenicity data to understand immune responses and predict immunogenicity of MHC-binding peptides accurately. RESULTS: This study proposes a computational method to mine a feature set of informative physicochemical properties from MHC class I binding peptides to design a support vector machine (SVM) based system (named POPI) for the prediction of peptide immunogenicity. High performance of POPI arises mainly from an inheritable bi-objective genetic algorithm, which aims to automatically determine the best number m out of 531 physicochemical properties, identify these m properties and tune SVM parameters simultaneously. The dataset consisting of 428 human MHC class I binding peptides belonging to four classes of immunogenicity was established from MHCPEP, a database of MHC-binding peptides (Brusic et al., 1998). POPI, utilizing the m = 23 selected properties, performs well with the accuracy of 64.72% using leave-one-out cross-validation, compared with two sequence alignment-based prediction methods ALIGN (54.91%) and PSI-BLAST (53.23%). POPI is the first computational system for prediction of peptide immunogenicity based on physicochemical properties. AVAILABILITY: A web server for prediction of peptide immunogenicity (POPI) and the used dataset of MHC class I binding peptides (PEPMHCI) are available at http://iclab.life.nctu.edu.tw/POPI     

Prediction of liquid chromatographic retention times of peptides generated by protease digestion of the Escherichia coli proteome using artificial neural networks

We developed a computational method to predict the retention times of peptides in HPLC using artificial neural networks (ANN). We performed stepwise multiple linear regressions and selected for ANN input amino acids that significantly affected the LC retention time. Unlike conventional linear models, the trained ANN accurately predicted the retention time of peptides containing up to 50 amino acid residues. In 834 peptides, there was a strong correlation (R2 = 0.928) between measured and predicted retention times. We demonstrated the utility of our method by the prediction of the retention time of 121,273 peptides resulting from LysC-digestion of the Escherichia coli proteome. Our approach is useful for the proteome-wide characterization of peptides and the identification of unknown peptide peaks obtained in proteome analysis.     

Physics and chemistry-driven artificial neural network for predicting bioactivity of peptides and proteins and their design

Predicting the bioactivity of peptides and proteins is an important challenge in drug development and protein engineering. In this study we introduce a novel approach, the so-called “physics and chemistry-driven artificial neural network (Phys-Chem ANN)”, to deal with such a problem. Unlike the existing ANN approaches, which were designed under the inspiration of biological neural system, the Phys-Chem ANN approach is based on the physical and chemical principles, as well as the structural features of proteins. In the Phys-Chem ANN model the “hidden layers” are no longer virtual “neurons”, but real structural units of proteins and peptides. It is a hybridization approach, which combines the linear free energy concept of quantitative structure-activity relationship (QSAR) with the advanced mathematical technique of ANN. The Phys-Chem ANN approach has adopted an iterative and feedback procedure, incorporating both machine-learning and artificial intelligence capabilities. In addition to making more accurate predictions for the bioactivities of proteins and peptides than is possible with the traditional QSAR approach, the Phys-Chem ANN approach can also provide more insights about the relationship between bioactivities and the structures involved than the ANN approach does. As an example of the application of the Phys-Chem ANN approach, a predictive model for the conformational stability of human lysozyme is presented.     

Modeling an active conformation for linear peptides and design of a competitive inhibitor for HMG-CoA reductase

This study presents an approach that can be used to search for lead peptide candidates, including unconstrained structures in a recognized sequence. This approach was performed using the design of a competitive inhibitor for 3-hydroxy-3-methylglutaryl CoA reductase (HMGR). In a previous design for constrained peptides, a head-to-tail cyclic structure of peptide was used as a model of linear analog in searches for lead peptides with a structure close to an active conformation. Analysis of the conformational space occupied by the peptides suggests that an analogical approach can be applied for finding a lead peptide with an unconstrained structure in a recognized sequence via modeling a cycle using fixed residues of the peptide backbone. Using the space obtained by an analysis of the bioactive conformations of statins, eight cyclic peptides were selected for a peptide library based on the YVAE sequence as a recognized motif. For each cycle, the four models were assessed according to the design criterion ("V" parameter) applied for constrained peptides. Three cyclic peptides (FGYVAE, FPYVAE, and FFYVAE) were selected as lead cycles from the library. The linear FGYVAE peptide (IC(50) = 0.4 microM) showed a 1200-fold increase the inhibitory activity compared to the first isolated LPYP peptide (IC(50) = 484 microM) from soybean. Experimental analysis of the modeled peptide structures confirms the appropriateness of the proposed approach for the modeling of active conformations of peptides.     

What are artificial neural networks and what they can do?

The artificial neural networks (ANN) are very often applied in many areas of toxicology for the solving of complex problems, such as the prediction of chemical compound properties and quantitative structure-activity relationship. The aim of this contribution is to give the basic knowledge about conception of ANN, theirs division and finally, the typical application of ANN will be discussed. Due to the diversity of architectures and adaptation algorithms, the ANNs are used in the broad spectrum of applications from the environmental processes modeling, through the optimization to quantitative structure-activity relationship (QSAR) methods. In addition, especially ANNs with Kohonen learning are very effective classification tool. The ANNs are mostly applied in cases, where the commonly used methods does not work.