Kann KI ein neues Bakterium erfinden, das ein lebensrettendes Medikament produziert ?

Der Schnittpunkt von KI und synthetischer Biologie hat neue Möglichkeiten in der Arzneimittelforschung erschlossen. Forschende nutzten kürzlich KI, um neuartige Enzyme und Stoffwechselwege zu entwerfen, wodurch die Erschaffung völlig neuer Organismen möglich wurde. Diese gezüchteten Bakterien können Verbindungen produzieren, die in der Natur nicht vorkommen. Dieser Durchbruch stellt die traditionellen Grenzen sowohl der Biologie als auch der Technologie infrage.

Background

The intersection of AI and synthetic biology has catalyzed advances in drug discovery, enabling the design of novel enzymes and metabolic pathways to create organisms that produce compounds beyond natural occurrence. Researchers have engineered bacteria to synthesize complex molecules such as insulin and antibiotics by incorporating synthetic gene circuits and optimizing metabolic pathways. However, constructing a de novo bacterial strain from scratch that reliably manufactures a previously unknown life-saving drug remains unachieved as of mid-2024. This challenge stems from the need for predictive models encompassing genetics, metabolism, and ecological interactions—capabilities beyond current bioengineering tools like CRISPR and automated DNA synthesis. Recent progress in AI-driven protein design tools such as RFdiffusion and AlphaFold3, alongside synthetic biology platforms like BioBrick and CRISPR libraries, facilitates targeted modifications, yet no group has publicly reported a bacterial species capable of synthesizing a novel, unnatural therapeutic compound with proven human efficacy and safety. While lab-evolved organisms like E. coli Nissle 1917 show therapeutic delivery potential, full de novo drug synthesis at clinical efficacy levels remains experimental. These limitations highlight the complex interplay between computational prediction, genetic engineering, and biochemical validation in advancing synthetic biology towards revolutionary medical applications.