Deep Learning Empowers the Discovery of Self-Assembling Peptides with Over 10 Trillion Sequences

Self-assembling of peptides is essential for a variety of biological and medical applications. However, it is challenging to investigate the self-assembling properties of peptides within the complete sequence space due to the enormous sequence quantities. Here, it is demonstrated that a transformer-based deep learning model is effective in predicting the aggregation propensity (AP) of peptide systems, even for decapeptide and mixed-pentapeptide systems with over 10 trillion sequence quantities. Based on the predicted AP values, not only the aggregation laws for designing self-assembling peptides are derived, but the transferability relation among the APs of pentapeptides, decapeptides, and mixed pentapeptides is also revealed, leading to discoveries of self-assembling peptides by concatenating or mixing, as consolidated by experiments. This deep learning approach enables speedy, accurate, and thorough search and design of self-assembling peptides within the complete sequence space of oligopeptides, advancing peptide science by inspiring new biological and medical applications. Peptide self-assembly is essential for a variety of applications in biological and medical sciences. Transformer-based deep learning has significantly expanded the explorable peptide sequence space to over tens of trillions of sequences. Design rules for achieving assemblies by mono-component peptides, as well as by concatenation or mixing of peptides are elucidated, enabling fast, accurate, and thorough design.image

Automated summary

This study demonstrates the effectiveness of a transformer-based deep learning model in predicting the aggregation propensity of peptide systems, even with enormous sequence quantities. Design rules for self-assembling peptides are derived based on predicted values, and the transferability relation among different peptides is revealed, leading to discoveries of self-assembling peptides through concatenation or mixing. This deep learning approach enables fast, accurate, and thorough design of self-assembling peptides within the complete sequence space, advancing peptide science with new applications.