# Revolutionizing microbial treasure troves: innovative strategies for natural products discovery

## TL;DR

This review explores new methods for discovering valuable microbial metabolites using innovative strategies like OSMAC and machine learning to boost drug development.

## Contribution

The paper introduces a comprehensive framework integrating OSMAC and genomics with machine learning for efficient microbial natural product discovery.

## Key findings

- OSMAC and genomics-driven approaches are overcoming cultivation limitations in microbial metabolite discovery.
- Machine learning is enabling faster identification of bioactive compounds with novel structures and mechanisms.
- Case studies from 2019 to 2025 highlight newly discovered compounds with potential for antibiotics and therapeutics.

## Abstract

Microorganisms represent Earth's most abundant biological resource, producing metabolites of immense value across medicine, agriculture, and industry. Conventional cultivation and screening techniques, however, suffer from inefficiency and fail to meet contemporary demands. Providing a comprehensive overview, this review details how the One Strain Many Compounds (OSMAC) strategy—addressing cultivation bottlenecks—and genomics-driven mining approaches are revolutionizing the discovery of novel microbial metabolites. Crucially, it underscores the broad adoption of innovative technologies like machine learning to enable faster, more effective gene and structure targeting. Synthesizing case studies from 2019 to 2025, the review catalogs newly identified compounds and their bioactivities, while outlining future research directions to establish a theoretical framework for efficient microbial natural product exploration. These advanced discovery strategies are significantly accelerating the identification of structurally diverse lead compounds with novel mechanisms of action, thereby revitalizing pipelines for new antibiotic, anticancer, and therapeutic drug development.

The online version contains supplementary material available at 10.1007/s13659-025-00565-0.