AI identifies antimicrobial peptides to target drug-resistant bacteria

Author: Aoife Doherty

Last year, the World Health Organization declared antimicrobial resistance to be a global health emergency. We are not developing novel, effective antimicrobial drugs quickly enough. In turn, the success of procedures such as major surgery and cancer chemotherapy are becoming severely compromised by antibiotic-resistant bacteria such as MRSA, and drug-resistant versions of life threatening diseases including TB and malaria are emerging and have the potential to kill millions. Therefore, we need to find novel, effective antimicrobial molecules fast.

The search for more powerful antibiotics has led scientists to investigate peptides; small, naturally occurring molecules that have effectively treated a range of common health issues such as diabetes and hypertension. In a bid to speed up the laborious, time-consuming, expensive and often fruitless expedition to develop drugs, researchers from Massachusetts Institute of Technology (MIT) and the Catholic University of Brasilia have used AI (a range of computational algorithms) to take a natural plant peptide that is known to have weak antibiotic activity, and to generate variants of the original molecule that are likely to have potent antimicrobial activity. Some of the most promising candidates that Porto et al. identified using this approach were tested as antibiotics in the lab, and one molecule in particular, guavanin 2, was shown to be effective, particularly against the ‘gram-negative’ family of bacteria that includes many of the bacterial species that are responsible for some of the most common hospital-acquired infections, including pneumonia. Their findings are the latest to demonstrate that we can use computational approaches to explore natural products, which could be used to design effective peptide antibiotics.

The aim of Porto et al.’s research reflects the ethos of Nuritas™. At Nuritas™, we are making an incredible amount of life-changing discoveries by successfully combining biological information about naturally occurring peptides with the most advanced computational techniques to positively affect the health of billions of people.

For further information, the full article is available here.

Reference: Porto WF, Irazazabal L, Alves ESF, Ribeiro SM, Matos CO, Pires ÁS, Fensterseifer ICM, Miranda VJ, Haney EF, Humblot V, Torres MDT, Hancock REW, Liao LM, Ladram A, Lu TK, de la Fuente-Nunez C, Franco OL.  (2018) “In silico optimization of a guava antimicrobial peptide enables combinatorial exploration for peptide design.” Nature communications 9(1): 1490.