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

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.

Identify Bitter Peptides by Using Deep Representation Learning Features

“…We compared the predictive performance of iBitter-DRLF with existing methods, including iBitter-Fuse [ 18 ], MIMML [ 20 ], iBitter-SCM [ 17 ], and BERT4Bitter [ 19 ] to assess the effectiveness and utility of our method against its competitors. Independent test results for iBitter-DRLF and the existing methods are compared in Table 4 .…”

“…A number of computational methods based on quantitative structure-activity relationship modeling (QSAR) and machine learning have been developed to predict the bitterness of polypeptides [ 9 , 10 , 11 , 12 , 13 , 14 ]. For example, BitterX [ 15 ], BitterPredict [ 16 ], iBitter-SCM [ 17 ], and iBitter-Fuse [ 18 ] using traditional sequence features were proposed to identify bitter peptides and showed increasing performance. Now the methods for identifying bitter peptides sequences are focused on feature engineering.…”

A survey on computational taste predictors

“…To enhance model performance and to increase the understanding of the model, a feature selection was normally applied, such as the V-WSP unsupervised variable reduction method and genetic algorithm-based technique [ 29 , 51 , 56 ], feature importance obtained from the RF [ 49 , 54 ], the Boruta algorithm, and the PCA [ 59 ]. However, none of these approaches consider the multi-objective nature of the dimensionality reduction techniques and thus fail to balance between the objectives of optimizing prediction performance measured in multiple classifications and regression metrics, minimizing the number of selected features and maximizing the overall interpretability/explainability of the derived prediction models.…”

“…After the development of iBitter-SCM and BERT4Bitter, the same research group implemented an improved bitter/non-bitter peptides predictor, called iBitter-Fuse [ 56 ]. This model overcomes some of the main drawbacks of the previous ones, including the generalization capability linked to the feature representation, overfitting, redundancy, and the overall performance.…”

Editorial of Special Issue “Deep Learning and Machine Learning in Bioinformatics”

“…Campos et al employed machine learning approaches to identify essential genes common to both Caenorhabditis elegans and Drosophila melanogaster [ 15 ]. Charoenkwan et al developed a sequence-based predictor, named iBitter-Fuse, to identify bitter peptides by fusing multi-view features [ 16 ]. Jabeen et al adopted a random forest model to identify novel high activity agonists of human ectopic olfactory receptors [ 17 ].…”