Prophages are phages that are integrated into bacterial genomes and which are key to understanding many aspects of bacterial biology. Their extreme diversity means they are challenging to detect using sequence similarity, yet this remains the paradigm and thus many phages remain unidentified. We present a novel, fast and generalizing machine learning method based on feature space to facilitate novel prophage discovery. To validate the approach, we reanalyzed publicly available marine viromes and single-cell genomes using our feature-based approaches and found consistently more phages than were detected using current state-of-the-art tools while being notably faster. This demonstrates that our approach significantly enhances bacteriophage discovery and thus provides a new starting point for exploring new biologies.

Genes of unknown function are among the biggest challenges in molecular biology, especially in microbial systems, where 40%-60% of the predicted genes are unknown. Despite previous attempts, systematic approaches to include the unknown fraction into analytical workflows are still lacking. Here, we present a conceptual framework, its translation into the computational workflow AGNOSTOS and a demonstration on how we can bridge the known-unknown gap in genomes and metagenomes. By analyzing 415,971,742 genes predicted from 1,749 metagenomes and 28,941 bacterial and archaeal genomes, we quantify the extent of the unknown fraction, its diversity, and its relevance across multiple organisms and environments. The unknown sequence space is exceptionally diverse, phylogenetically more conserved than the known fraction and predominantly taxonomically restricted at the species level. From the 71M genes identified to be of unknown function, we compiled a collection of 283,874 lineage-specific genes of unknown function for Cand. Patescibacteria (also known as Candidate Phyla Radiation, CPR), which provides a significant resource to expand our understanding of their unusual biology. Finally, by identifying a target gene of unknown function for antibiotic resistance, we demonstrate how we can enable the generation of hypotheses that can be used to augment experimental data.

AbstractBridging the gap between the known and the unknown coding sequence space is one of the biggest challenges in molecular biology today. This challenge is especially extreme in microbiome analyses where between 40% to 60% of the coding sequences detected are of unknown function, and ignoring this fraction limits our understanding of microbial systems. Discarding the uncharacterized fraction is not an option anymore. Here, we present an in-depth exploration of the microbial unknown fraction through the lenses of a conceptual framework and a computational workflow we developed to unify the microbial known and unknown coding sequence space. Our approach partitions the coding sequence space in gene clusters and contextualizes them with genomic and environmental information. We analyzed 415,971,742 genes predicted from 1,749 metagenomes and 28,941 bacterial and archaeal genomes putting into perspective the extent of the unknown fraction, its diversity, and its relevance in a genomic and environmental context. With the identification of a target gene of unknown function for antibiotic resistance, we demonstrate how a contextualized unknown coding sequence space provides a robust framework for the generation of hypotheses that can be used to augment experimental data.

No abstract