Multi-label Learning for Predicting the Activities of Antimicrobial Peptides

Abstract

Antimicrobial peptides (AMPs) are peptide antibiotics with a broad spectrum of antimicrobial activities. Activity prediction of AMPs from their amino acid sequences is of great therapeutic importance but imposes challenges on prediction methods due to label interactions. In this paper we propose a novel multi-label learning model to address this problem. A weighted K-nearest neighbor classifier is adopted for efficient representation learning of the sequence data. A multiple linear regression model is then employed to learn a mapping from the classifier score vectors to the target labels, with label correlations considered. Several popular multi-label learning algorithms and feature extraction methods were tested on a comprehensive, up-to-date AMP dataset with twelve biological activities covered and its filtered version with five activities covered. The experimental results showed that our proposed method has competitive performance with previous works and could be used as a powerful engine for activity prediction of AMPs.

Accelerating antibiotic discovery through artificial intelligence

“…The in silico identification of putative proteolytic cleavage sites can inform AMP lead selection and guide sequence optimization for increased stability. Cleavage site prediction has been explored through the lens of drug development 69 and other protein informatics applications using classification and regression mode SVM [70][71][72][73][74] , convolutional neural network 75 , conditional random field classifier 76 , and logistic regression models 77 . Similarly, the stability of drug-like chemicals has been modeled using an attention-based graph convolution neural network 78 and Naive Bayes classifier 79 .…”

Computational Methods and Tools in Antimicrobial Peptide Research

“…To conclude, the field of ML-assisted prediction of AMP sequences and activities is an ever-expanding one, augmented by the large number of available methods and models. These methods are not mutually exclusive; in fact, they are often used in conjunction to bolster each other and improve the quality of the results. , Simpler statistical tools are also available, such as logistic regression . These methods can be used in many forms and are available as open access libraries of scripts in packages such as scikit-learn for python or the mlr package for R; amPEPpy, for example, employs an RF classifier using scikit-learn, and IAMPE implemented five machine learning models (NB, KNN, SVM, RF, and XGBoost) with scikit-learn.…”

“…109,153 Simpler statistical tools are also available, such as logistic regression. 154 These methods can be used in many forms and are available as open access libraries of scripts in packages such as scikit-learn for python 155 or the mlr package for R; 156 amPEPpy, for example, employs an RF classifier using scikit-learn, 116 and IAMPE implemented five machine learning models (NB, KNN, SVM, RF, and XGBoost) with scikit-learn. As our computational capabilities improve and we can perform more calculations faster, machine-aided prediction will play an increasing role in AMP research.…”

The Road from Host-Defense Peptides to a New Generation of Antimicrobial Drugs

“…The AMPs were subjected to in silico analysis of their antimicrobial activity using several bioinformatic tools [ 21 , 81 , 82 , 83 , 84 , 85 , 86 , 87 ]. For example, the online tool APD3 calculates the different parameters related to the possible antimicrobial activity of the peptides (e.g., net charge, length, percentage of hydrophobic residues, helicity) [ 81 , 82 ].…”