IBitter-Fuse: A Novel Sequence-Based Bitter Peptide Predictor by Fusing Multi-View Features.

Phasit Charoenkwan,Chanin Nantasenamat,Pietro Lio’,Md Mehedi Hasan,Mohammad Ali Moni,Watshara Shoombuatong

doi:10.3390/ijms22168958

Abstract

Accurate identification of bitter peptides is of great importance for better understanding their biochemical and biophysical properties. To date, machine learning-based methods have become effective approaches for providing a good avenue for identifying potential bitter peptides from large-scale protein datasets. Although few machine learning-based predictors have been developed for identifying the bitterness of peptides, their prediction performances could be improved. In this study, we developed a new predictor (named iBitter-Fuse) for achieving more accurate identification of bitter peptides. In the proposed iBitter-Fuse, we have integrated a variety of feature encoding schemes for providing sufficient information from different aspects, namely consisting of compositional information and physicochemical properties. To enhance the predictive performance, the customized genetic algorithm utilizing self-assessment-report (GA-SAR) was employed for identifying informative features followed by inputting optimal ones into a support vector machine (SVM)-based classifier for developing the final model (iBitter-Fuse). Benchmarking experiments based on both 10-fold cross-validation and independent tests indicated that the iBitter-Fuse was able to achieve more accurate performance as compared to state-of-the-art methods. To facilitate the high-throughput identification of bitter peptides, the iBitter-Fuse web server was established and made freely available online. It is anticipated that the iBitter-Fuse will be a useful tool for aiding the discovery and de novo design of bitter peptides.

Highlights

To protect themselves from environmental toxins, mammalian species, including humans, are averse to bitter-tasting substances. [1]
Several computational methods based on quantitative structure– activity relationship (QSAR) modeling have been published on the prediction of peptide bitterness [10,11,12,13,14,15]
In BitterX, sequential minimal optimization (SMO), logistic regression (LR) and random forest (RF) were employed to develop ML-based models in order to discriminate bitter from non-bitter compounds

Summary

Introduction

To protect themselves from environmental toxins, mammalian species, including humans, are averse to bitter-tasting substances. [1]. Experimental methods are considered to be reliable approaches for characterizing the bitterness of peptides [5,8,9], they are usually time-consuming and expensive. Due to their convenience and high efficiency, machine-learning (ML) methods have attracted increasing attention in the field of bioinformatics. In BitterX, sequential minimal optimization (SMO), logistic regression (LR) and random forest (RF) were employed to develop ML-based models in order to discriminate bitter from non-bitter compounds. In their experimental setting, training (70%) and hold-out test (30%) datasets were constructed for model development and validation. BitterPredict was developed by Dagan-Wiener et al [16] in order to identify bitter compounds based on the information of chemical structures

Methods

Results

Conclusion