The Use of DeepQSAR Models for The Discovery of Peptides with Enhanced Antimicrobial and Antibiofilm Potential

Increasing concerns regarding prolonged antibiotic usage have spurred the search for alternative treatments. Antimicrobial peptides (AMPs), first discovered in the 1980s, have exhibited significant potential against a broad range of bacteria [1]. Short-sequenced AMPs are abundant in nature and present across various organisms [2]. With lower likelihood of developing resistance, thousands of cationic and hydrophobic AMPs have been identified to act selectively on microbes, making them excellent anti-infective candidates in medicine, food, farming, and agriculture. Recently, machine learning technologies such as Quantitative Structure Activity Relationships (QSAR) have enabled expedited discovery of potential AMPs with broad-spectrum antibacterial activity as the amount of available AMP training data increases [3]. Among those, Deep QSAR has recently emerged as a distinct type of application that utilizes conventional molecular descriptors [4] in conjunction with more powerful deep learning (DL) models [5]. Here, we demonstrate the power of Deep QSAR in predicting broad-spectrum AMP activity. Our approach utilizes a recurrent neural network to perform QSAR regression modeling and achieves nearly 90% five-fold cross-validated accuracy in predicting AMP antimicrobial and antibiofilm potential. Using the developed approach, we designed 100 novel peptides, of which 44 experimentally demonstrated more effective antibiofilm activity and 31 peptides exhibited stronger antimicrobial activity compared to the known antibiofilm/antimicrobial peptide IDR-1018 [6] used as a positive control. Notably, 29 of those peptides demonstrated improvements of both antimicrobial and, particularly, antibiofilm properties, making them suitable prototypes for pre-clinical development and demonstrating efficacy of DeepQSAR modeling in identifying novel and potent AMPs.