Abstract
Activity cliff (AC) is a phenomenon that a pair of similar molecules differby a small structural alternation but exhibit a large difference in theirbiochemical activities. The AC of small molecules has been extensivelyinvestigated but limited knowledge is accumulated about the AC phenomenon inpeptides with canonical amino acids. This study introduces a quantitativedefinition and benchmarking framework AMPCliff for the AC phenomenon inantimicrobial peptides (AMPs) composed by canonical amino acids. Acomprehensive analysis of the existing AMP dataset reveals a significantprevalence of AC within AMPs. AMPCliff quantifies the activities of AMPs by themetric minimum inhibitory concentration (MIC), and defines 0.9 as the minimumthreshold for the normalized BLOSUM62 similarity score between a pair ofaligned peptides with at least two-fold MIC changes. This study establishes abenchmark dataset of paired AMPs in Staphylococcus aureus from the publiclyavailable AMP dataset GRAMPA, and conducts a rigorous procedure to evaluatevarious AMP AC prediction models, including nine machine learning, four deeplearning algorithms, four masked language models, and four generative languagemodels. Our analysis reveals that these models are capable of detecting AMP ACevents and the pre-trained protein language ESM2 model demonstrates superiorperformance across the evaluations. The predictive performance of AMP activitycliffs remains to be further improved, considering that ESM2 with 33 layersonly achieves the Spearman correlation coefficient=0.50 for the regression taskof the MIC values on the benchmark dataset. Source code and additionalresources are available at https://www.healthinformaticslab.org/supp/ orhttps://github.com/Kewei2023/AMPCliff-generation.