Recurrent Neural Network Model for Constructive Peptide Design

Abstract

We present a generative long short-term memory (LSTM) recurrent neural network (RNN) for combinatorial de novo peptide design. RNN models capture patterns in sequential data and generate new data instances from the learned context. Amino acid sequences represent a suitable input for these machine-learning models. Generative models trained on peptide sequences could therefore facilitate the design of bespoke peptide libraries. We trained RNNs with LSTM units on pattern recognition of helical antimicrobial peptides and used the resulting model for de novo sequence generation. Of these sequences, 82% were predicted to be active antimicrobial peptides compared to 65% of randomly sampled sequences with the same amino acid distribution as the training set. The generated sequences also lie closer to the training data than manually designed amphipathic helices. The results of this study showcase the ability of LSTM RNNs to construct new amino acid sequences within the applicability domain of the model and motivate their prospective application to peptide and protein design without the need for the exhaustive enumeration of sequence libraries.

Designing Anticancer Peptides by Constructive Machine Learning

“…RNNs are suited to address sequence‐based representations of molecules, such as molecular SMILES strings and amino acid sequences. Recently, we have proposed an RNN strategy with long short‐term memory (LSTM) cells for sequence‐based peptide design . Here, we apply constructive RNNs to automatically generate novel membranolytic anticancer peptides (ACPs) and investigate the prospective applicability of this novel concept in peptide design.…”

“…The computational approach used was based on an RNN model with LSTM cells and consisted of two steps (Figure ). In the first step, we developed a generic model that learned the grammar of 10 000 presumably α‐helical cationic amphipathic peptides represented as amino acid sequences in one‐letter code.…”

“…[9] Here, we apply constructive RNNs to automatically generate novel membranolytic anticancer peptides (ACPs) and investigate the prospective applicabilityo ft his novel concept in peptide design.T he types of ACPs considered kill targetc ancercells by non-enzymatic membrane lysis. [10] The computational approach used was based on an RNN modelw ith LSTM cells [9] and consisted of two steps (Figure 1). In the first step, we developed ag eneric model that learned the grammar of 10 000 presumably a-helical cationic amphipathicp eptides represented as amino acid sequences in oneletter code.…”

Prediction of the Health Effects of Food Peptides and Elucidation of the Mode‐of‐action Using Multi‐task Graph Convolutional Neural Network

“…For example, peptides derived from bovine casein have antibacterial activity, [3] peptides derived from milk and sardines have the effects of lowering blood pressure, [4,5] peptides derived from soybeans have the effect of lowering cholesterol, [6] peptides derived from pork have antithrombotic activity, [7] and peptides derived from hoki have antioxidant activity. [10][11][12] There are no previous studies on food peptides from computational viewpoints. [9] Bioactive food peptides can be a lead compound for the development of new drugs and functional foods.…”

“…[19] Examples of the ML methods include partial least square, [20] least absolute shrinkage and selection operator, [21] support vector machine, [22] random forest, [23] and gradient boosting decision tree. [25,26] For their applications to peptides, the recurrent neural network (RNN) has been adapted to generate novel peptide sequences, [27] and the convolutional neural network (CNN) has been used to predict interactions between peptides and the major histocompatibility com-plex (MHC). [25,26] For their applications to peptides, the recurrent neural network (RNN) has been adapted to generate novel peptide sequences, [27] and the convolutional neural network (CNN) has been used to predict interactions between peptides and the major histocompatibility com-plex (MHC).…”