Molecular arms race classifier for decrypting venom peptide and ion channel interactions

Ion channels are essential regulators of cellular processes, and their dysfunction underpins a wide range of diseases. Peptides, particularly those derived from venom, have proven to be powerful modulators of ion channel activity, with several already in clinical use as FDA-approved therapeutics. Despite the remarkable potential of venom peptides and proteins, their interactions with ion channels remain largely uncharacterized. In this study, we introduce a machine learning approach termed Molecular Arms Race Classifier (MARC), which predicts the molecular targets – specifically sodium, potassium, and calcium ion channels – of cysteine-rich venom peptides. The MARC model was trained, cross-validated, and tested on 5,165 peptide and protein sequences sourced from in-house venom gland transcriptomes and public databases spanning diverse taxa, including sea anemones, snakes, scorpions, spiders, cone and terebrid snails. As a machine learning-based classifier of ion channel specificities, the MARC model captures the venom peptides' molecular features consisting of structural and sequence information with ESM and ProtLearn. Key findings from our analysis include the identification of 29 novel terebrid snail venom peptides, teretoxins, predicted to target potassium ion channels. Notably, most predicted potassium channel modulators (19 of 29) were classified with a distinct venom peptide cysteine framework, Framework 9. These predictions were further tested using molecular docking and dynamics simulations, validating the robustness of our MARC algorithm. MARC is a cost-effective solution for screening vast peptide and protein libraries and identifying ion channel targeting compound candidates. This work deepens our understanding of the structural and functional features of venom peptides and proteins (or venom compounds) that selectively target different ion channel classes.