AMPpred-MFA: An Interpretable Antimicrobial Peptide Predictor with a Stacking Architecture, Multiple Features, and Multihead Attention

Abstract

Antimicrobial peptides (AMPs) are small molecular polypeptides that can be widely used in the prevention and treatment of microbial infections. Although many computational models have been proposed to help identify AMPs, a high-performance and interpretable model is still lacking. In this study, new benchmark data sets are collected and processed, and a stacking deep architecture named AMPpred-MFA is carefully designed to discover and identify AMPs. Multiple features and a multihead attention mechanism are utilized on the basis of a bidirectional long short-term memory (LSTM) network and a convolutional neural network (CNN). The effectiveness of AMPpred-MFA is verified through five independent tests conducted in batches. Experimental results show that AMPpred-MFA achieves a state-of-the-art performance. The visualization interpretability analyses and ablation experiments offer a further understanding of the model behavior and performance, validating the importance of our feature representation and stacking architecture, especially the multihead attention mechanism. Therefore, AMPpred-MFA can be considered a reliable and efficient approach to understanding and predicting AMPs.

deepAMPNet: a novel antimicrobial peptide predictor employing AlphaFold2 predicted structures and a bi-directional long short-term memory protein language model

“…Utilizing the XUAMP dataset proposed by Xu et al (2021) we compared deepAMPNet with 11 state-of-the-art AMPs predictors including: amPEPpy ( Lawrence et al, 2021 ), AMPfun ( Chung et al, 2019 ), AMPEP ( Bhadra et al, 2018 ), ADAM-HMM ( Lee et al, 2015 ), ampir ( Fingerhut et al, 2021 ), AMPScannerV2 ( Veltri, Kamath & Shehu, 2018 ), AMPGram ( Burdukiewicz et al, 2020 ), Deep-AMPEP30 ( Yan et al, 2020 ), CAMP-ANN ( Waghu et al, 2016 ), sAMPpred-GAT ( Yan et al, 2023 ) and AMPpred-MFA ( Li et al, 2023 ). To avoid overestimating the performance of deepAMPNet, sequences with similarity above 90% with the XUAMP dataset were removed from the deepAMPNet training dataset.…”

“…We comprehensively evaluated the performance of deepAMPNet on the MFA_test dataset with other 11 state-of-the-art AMPs predictors including: ampir ( Fingerhut et al, 2021 ), CAMP3 (RF), CAMP3 (SVM), CAMP3 (DA), CAMP3 (ANN) ( Waghu et al, 2016 ), iAMPpred ( Meher et al, 2017 ), AMPScannerV2 ( Veltri, Kamath & Shehu, 2018 ), AI4AMP ( Lin et al, 2021 ), AMPlify ( Li et al, 2022 ), AMPfun ( Chung et al, 2019 ) and AMPpred-MFA ( Li et al, 2023 ). We present ROC curves of all predictors in Fig.…”

“…The non-antimicrobial peptides (Non-AMPs) was reaped from Uniprot, processed similarly, and randomly chosen, resulting in a balance of 1,536 AMPs and 1,536 Non-AMPs. MFA_test, presented by Li et al (2023) , was similarly treated and contained 615 AMPs and 606 Non-AMPs.…”

Editorial: Machine Learning in Bio-cheminformatics

“…( a ) protein structure prediction and functional annotation . A variety of intelligent frameworks were introduced for predicting the structures of blood-brain barrier-penetrating peptide, antimicrobial peptide, and intrinsically disordered protein. − Elia Venanzi et al applied a machine learning method that integrated protein structure, sequence, and dynamics to predict the enzymatic activity of bovine intestinal kinase variants. Zachary Smith et al identified druggable binding sites of protein target using graph neural networks with attention.…”