Classifying antimicrobial and multifunctional peptides with Bayesian network models

Abstract

Bayesian network models are finding success in characterizing enzyme‐catalyzed reactions, slow conformational changes, predicting enzyme inhibition, and genomics. In this work, we apply them to statistical modeling of peptides by simultaneously identifying amino acid sequence motifs and using a motif‐based model to clarify the role motifs may play in antimicrobial activity. We construct models of increasing sophistication, demonstrating how chemical knowledge of a peptide system may be embedded without requiring new derivation of model fitting equations after changing model structure. These models are used to construct classifiers with good performance (94% accuracy, Matthews correlation coefficient of 0.87) at predicting antimicrobial activity in peptides, while at the same time being built of interpretable parameters. We demonstrate use of these models to identify peptides that are potentially both antimicrobial and antifouling, and show that the background distribution of amino acids could play a greater role in activity than sequence motifs do. This provides an advancement in the type of peptide activity modeling that can be done and the ease in which models can be constructed.

Investigating Active Learning and Meta-Learning for Iterative Peptide Design

“…The task model is a convolutional neural network whose structure was partially based on that of a model from our past work in peptide modeling. 81 The model structure is shown schematically in Figure 1 . The first layer of the neural network is a convolutional layer with a weight matrix of dimension [ W × A × K ], where W and K conceptually represent peptide “motif widths” and number of “motif classes,” respectively, and A is the length of the amino acid alphabet considered (20 for the naturally occurring peptides used here).…”

Encodings and models for antimicrobial peptide classification for multi-resistant pathogens

“…In addition, Table 5 summarizes the presented kernels. However, besides the popular algorithms mentioned above, further methods leveraged partial least squares [82, 83, 107], hidden Markov models [108], logistic regression [109] and Bayesian networks [110]. Furthermore, ensembles of several classifiers have been also successfully implemented, such as in [17] or [21], whereby often one classifier is trained with a particular sequence or structural encoding.…”

AMPGAN v2: Machine Learning Guided Design of Antimicrobial Peptides

“…Machine learning has been used to aid in the discovery and development of AMPs, with increasing adoption as both fields have matured. A large body of work has been developed based on predictive modeling of AMP properties, [18][19][20][21][22][23][24][25][26][27] and is generally labelled as quantitative structure-activity relationship (QSAR) modeling. The basic QSAR approach is to select a property of interest (e.g.…”

“…Many recent approaches to AMP discovery have relied on predictive models, [18][19][20][21][22][23][24][25][26][27] using a naive candidate generation method and a rejection sampling approach to identify promising candidates. Since the candidate generation method is naive, often many samples must be drawn before a promising peptide is found.…”

“…Machine learning has aided in the discovery and development of AMPs, with many recent approaches relying on predictive models. [18][19][20][21][22][23][24][25][26][27] Such approaches are usually labelled as quantitative structure-activity relationship (QSAR) models. The basic QSAR recipe is to select a property of interest (e.g., antimicrobial activity), train a machine learning model to predict that property using relatively easily obtained features (e.g., primary peptide structure), then apply the trained model to unlabelled samples to estimate the property of interest.…”