AI在古菌中识别出抗菌化合物

Core Insights - A research team from the University of Pennsylvania has utilized artificial intelligence (AI) to identify previously unknown antibacterial compounds from ancient microorganisms, specifically archaea, which have survived in extreme environments for billions of years [2][3] - The study highlights the potential of archaea as a novel molecular resource for antibiotic development, with compounds that may operate through mechanisms different from existing drugs [2] - The upgraded APEX tool was employed to predict which peptides from archaea could inhibit bacterial growth, resulting in the identification of over 12,000 candidate molecules from 233 archaea proteins [2] Summary by Sections - Research Methodology - The APEX AI model was retrained using thousands of additional peptides and information on pathogenic bacteria to enhance its predictive capabilities [2] - The team screened 233 archaea proteins and identified over 12,000 candidate molecules, which were found to differ from known antibacterial peptides, particularly in charge distribution [2] - Experimental Results - Out of the selected 80 candidate molecules, 93% showed effectiveness against at least one type of drug-resistant pathogen [3] - In animal tests, one of the three compounds demonstrated an ability to inhibit drug-resistant bacteria comparable to the commonly used last-resort antibiotic, polymyxin B [3] - Unlike most antibacterial peptides that disrupt bacterial outer membranes, these candidates appear to "cut off the power" from within the bacteria, disrupting their life-sustaining electrical signals [3]