Poster Presentation International Peptide Symposium 2023

Machine learning-guided discovery of antifungal peptides from hymenopteran venoms (#336)

J. Gabriel Osuna-Osuna 1 , Silvana Aguilar 2 , Mariela Marani 2 , Eugenio Mancera-Ramos 1 , Fabien Plisson 1
  1. Centre for Research and Advanced of the National Polytechnic Institute (CINVESTAV-IPN), Irapuato, GUANAJUATO, Mexico
  2. Patagonian Institute for the Study of Continental Ecosystems (IPEEC) - National Council for Scientific and Technical Research (CONICET), Puerto Madryn, Argentina

Candidiasis is a highly prevalent and severe infection caused by the pathogenic fungus Candida albicans. The increasing resistance to current antifungal therapies poses a significant public health challenge, emphasising the urgent need for novel antifungal agents.1 Venom peptides offer promise due to their diverse bioactivities, target specificity, and metabolic stability. Hymenopterans, including ants, bees, and wasps, represent an untapped source of biologically active peptides.2 However, accelerating the identification of peptide candidates while reducing hurdles to clinical trials is crucial in peptide-based drug discovery. Machine learning models provide a cost-effective and time-saving approach for identifying antimicrobial peptides from their primary sequences.3

In this study, we employed machine learning as a filtering system to assess the antifungal potential of several known peptides derived from hymenopteran venoms without prior biological profiles, predicting their antifungal and hemolytic activity. Five promising candidates were chemically synthesized using Solid Phase Peptide Synthesis and evaluated through microdilution assays against C. albicans. Our results revealed the antifungal activity of three synthetic peptides with varying minimum inhibitory concentrations, demonstrating the potential of hymenopteran venoms as a source of novel antifungal agents. This study highlights the power of machine learning in accelerating the discovery of potentially therapeutic compounds, leveraging machine learning to streamline the identification process and overcome traditional limitations in natural product drug discovery.

  1. Calderone, R.A. and Clancy, C.J. Candida and candidiasis. American Society for Microbiology Press. (2012) Online ISBN: 9781683670957, https://doi.org/10.1128/9781555817176.
  2. Guido-Patiño, J.C. and Plisson, F. Profiling hymenopteran venom toxins: Protein families, structural landscape, biological activities, and pharmacological benefits, Toxicon: X, 14, p. 100119. (2022) https://doi.org/10.1016/j.toxcx.2022.100119.
  3. Melo, M.C.R., Maasch, J.R.M.A. & de la Fuente-Nunez, C. Accelerating antibiotic discovery through artificial intelligence. Commun Biol 4, 1050 (2021). https://doi.org/10.1038/s42003-021-02586-0