TY - JOUR IS - 1 EP - 13 TI - Mining metagenomics data for novel bacterial nanocompartments SN - 2631-9268 AV - public KW - Science & Technology KW - Life Sciences & Biomedicine KW - Genetics & Heredity KW - Mathematical & Computational Biology KW - PROTEIN-STRUCTURE N1 - © The Author(s) 2024. Published by Oxford University Press on behalf of NAR Genomics and Bioinformatics. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. N2 - Encapsulin nanocompartments are prokaryotic protein-based organelles. T he y displa y div erse natural functions, including mineral storage and stress response. Encapsulins also ha v e applications in synthetic biology, drug deliv ery, v accines, and met abolic engineering . Disco v ering no v el encapsulins is challenging due to inconsistent annotations, and data contamination due to similarity with phage proteins. P re vious studies ha v e disco v ered thousands of encapsulin sequences from bacteria and archaea, but met agenomics dat abases were not specifically interrogated. Metagenomics can provide information on a much larger diversity of unculturable organisms and environmental samples than con v entional sequencing experiments, and metagenomic protein databases ha v e shed light on previously unexplored regions of the protein universe. This study le v erages de v elopments in deep learning for str uct ure and function prediction, to produce a dataset of o v er 1300 no v el putativ e encap- sulin sequences from the MGnify Protein Database. Some well-known encapsulins and their cargo proteins were identified, predominantly pero xidases and ferritin-lik e proteins. A potentially no v el encapsulin-associated biosynthetic gene cluster in v olv ed in producing cytoto xic or an- timicrobial saccharides was discovered using biosynthetic gene cluster prediction. Finally, a cluster of predicted str uct ures with no v el features not seen in experimentally solved encapsulin str uct ures was discovered using large-scale, deep learning-based str uct ure prediction of putative metagenomic encapsulins. UR - https://doi.org/10.1093/nargab/lqae025 ID - discovery10192201 Y1 - 2024/03/07/ A1 - Kashif-Khan, Naail A1 - Savva, Renos A1 - Frank, Stefanie JF - NAR Genomics and Bioinformatics PB - OXFORD UNIV PRESS VL - 6 ER -