Functional genomic signatures predict microbial culturability across the tree of life
Oduwole, I.; Babjac, A.; Royalty, T. M.; Hibbs, M.; Lloyd, K. G.; Emrich, S.; Steen, A. D.
Show abstract
Most microbial taxa on Earth remain uncultivated, limiting our ability to study their physiology, ecology, and roles in environmental processes. Although metagenome-assembled genomes (MAGs) have expanded access to uncultured phylogenetic diversity, the functional basis for culturability remains poorly understood. Here, we analyze the 52,515 MAGs from the Genomes from Earths Microbiomes (GEM) catalog to test two hypotheses: 1) genomes from uncultured microbes encode more functionally novel genes than those from cultured taxa, and 2) specific genomic features are systematically associated with culturability across phyla. To assess functional novelty, we aligned predicted proteins to SwissProt and measured sequence dissimilarity to the nearest curated homolog. We find that uncultured MAGs, particularly among Archaea, harbor substantially more divergent proteins. To identify genomic traits predictive of culturability, we combined pathway-level enrichment with LASSO regression and permutation-based feature importance. Cultured MAGs were consistently enriched in Clusters of Orthologous Groups (COG) pathways related to vitamin and cofactor biosynthesis (e.g., thiamine, folate, B12), energy metabolism (e.g., TCA cycle), and CRISPR-Cas systems--functions often depleted in uncultured counterparts. LASSO models identified a subset of these pathways as strong predictors of cultured status even in poorly sampled phyla, suggesting conserved genomic signatures of culturability. In contrast, pathways such as purine biosynthesis and NADH dehydrogenase were associated with uncultured lineages, highlighting potential barriers to cultivation. These results 1) demonstrate the great functional novelty of uncultured microbes, potentially offering unprecedented opportunities for discoveries of novel function, and 2) identify metabolic traits associated with culturability to inform future cultivation strategies. ImportanceThe vast majority of microbes are uncultured, which means they have never been characterized under laboratory conditions. We showed that genomic sequences of uncultured microbes have less similarity to characterized proteins compared to cultured microbes, revealing that there may be fundamental biological reasons why they are not cultured. We also showed that certain metabolic pathways, such as those related to vitamin and cofactor biosynthesis, can predict the ability of microbes to grow under laboratory conditions, and these pathways are abundant in highly cultured phyla, indicating how metabolic pathways can influence cultivation strategies.
Matching journals
The top 2 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Patterns of gene content and co-occurrence constrain the evolutionary path toward animal association in CPR bacteria 98%
- An ANI gap within bacterial species that advances the definitions of intra-species units 96%
- Single-strain behavior predicts responses to environmental pH and osmolality in the gut microbiota. 96%
Similar papers in this journal
Similar papers in this journal
- Toward quantifying the adaptive role of bacterial pangenomes during environmental perturbations 97%
- Halophilic microbial community compositional shift after a rare rainfall in the Atacama Desert 96%
- Differential global distribution of marine picocyanobacteria gene clusters reveals distinct niche-related adaptive strategies 95%
Similar papers in this journal
- TbasCO: Trait-based Comparative 'Omics Identifies Ecosystem-Level and Niche-Differentiating Adaptations of an Engineered Microbiome 96%
- Ecological constraints and evolutionary trade-offs shape nitrogen fixation across habitats 95%
- Culexarchaeia, a novel archaeal class of anaerobic generalists inhabiting geothermal environments 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.