Metagenomic Mining of Antimicrobial Compounds from Extreme Environments: A Systematic Review

Prayogo, Fitra Adi

Journal of Biomedical Sciences and Health
Vol. 3 No. 1 (2026)

Background: Antimicrobial resistance (AMR) is a major global health crisis, necessitating novel drug discovery approaches. Extreme environments harbor unique microbial communities that produce specialized metabolites, yet systematic assessment of their biosynthetic potential through metagenomics remains lacking. Objective: To systematically review evidence on metagenomic mining strategies for discovering biosynthetic gene clusters (BGCs) with antimicrobial potential from extreme environments. Methods: Following PRISMA 2020 guidelines, PubMed/MEDLINE, Web of Science, Scopus, and Google Scholar were searched through December 2025. The primary reviewer screened all 487 records; a blinded second reviewer independently verified a random 20% subset at each stage (κ = 0.79–0.85). Quality assessment used an adapted Newcastle-Ottawa Scale. Fifteen studies met all inclusion criteria. Results: The 15 included studies identified over 14,000 BGCs (excluding the Paoli et al. [2022] global ocean dataset reported separately) across Antarctic/psychrophilic (5 studies), marine/deep-sea (4 studies), halophilic/hypersaline (2 studies), arid/desert environments (2 studies), and extreme soil communities (2 studies). Dominant BGC classes included terpenes, NRPS, RiPPs, and PKS. Studies employing long-read sequencing (Oxford Nanopore/PacBio) recovered substantially more complete BGCs compared with short-read approaches. Between 60–99% of detected BGCs across most environments lacked characterized homologs in the MIBiG database. Experimental validation of predicted antimicrobial activity was limited: only 2 studies (13.3%) confirmed direct antimicrobial or cytotoxic bioactivity through bioassays or compound isolation; 1 additional study (6.7%) provided indirect evidence of active BGC expression via metatranscriptomics; and the remaining 12 studies (80%) relied solely on in silico prediction. Conclusion: Extreme-environment metagenomics reveals remarkable biosynthetic diversity with substantial novelty. Long-read sequencing and updated bioinformatic platforms have significantly enhanced BGC detection. The critical gap between computational prediction and experimental validation of antimicrobial bioactivity remains the primary barrier to therapeutic translation.