Search for: All records

Abstract MotivationHigh-throughput sequencing (HTS) is a modern sequencing technology used to profile microbiomes by sequencing thousands of short genomic fragments from the microorganisms within a given sample. This technology presents a unique opportunity for artificial intelligence to comprehend the underlying functional relationships of microbial communities. However, due to the unstructured nature of HTS data, nearly all computational models are limited to processing DNA sequences individually. This limitation causes them to miss out on key interactions between microorganisms, significantly hindering our understanding of how these interactions influence the microbial communities as a whole. Furthermore, most computational methods rely on post-processing of samples which could inadvertently introduce unintentional protocol-specific bias. ResultsAddressing these concerns, we present SetBERT, a robust pre-training methodology for creating generalized deep learning models for processing HTS data to produce contextualized embeddings and be fine-tuned for downstream tasks with explainable predictions. By leveraging sequence interactions, we show that SetBERT significantly outperforms other models in taxonomic classification with genus-level classification accuracy of 95%. Furthermore, we demonstrate that SetBERT is able to accurately explain its predictions autonomously by confirming the biological-relevance of taxa identified by the model. Availability and implementationAll source code is available at https://github.com/DLii-Research/setbert. SetBERT may be used through the q2-deepdna QIIME 2 plugin whose source code is available at https://github.com/DLii-Research/q2-deepdna. 

Abstract Restoring ecosystems requires the re-establishment of diverse soil microbial communities that drive critical ecosystem functions. In grasslands, restoration and management require the application of disturbances like fire and grazing. Disturbances can shape microbial taxonomic composition and potentially functional composition as well. We characterized taxonomic and functional gene composition of soil communities using whole genome shotgun metagenomic sequencing to determine how restored soil communities differed from pre-restoration agricultural soils and original remnant soils, how management affects soil microbes, and whether restoration and management affect the number of microbial genes associated with carbohydrate degradation. We found distinct differences in both taxonomic and functional diversity and composition among restored, remnant, and agricultural soils. Remnant soils had low taxonomic and functional richness and diversity, as well as distinct composition, indicating that restoration of agricultural soils does not re-create soil microbial communities that match remnants. Prescribed fire management increased functional diversity, which also was higher in more recently planted restorations. Finally, restored and post-fire soils included high abundances of genes encoding cellulose-degrading enzymes, so restorations and their ongoing management can potentially support functions important in carbon cycling. 

1. Ecosystem restoration often focuses on re‐establishing species richness and diversity of native organisms. However, effective restoration requires re‐establishment of ecosystem functions and processes by all trophic levels. Functional trait descriptions of communities, including decomposer communities, may provide more comprehensive evaluations of restoration activities and management than taxonomic community metrics alone. 2. We examined species and functional trait composition of dung beetle (Coleoptera: Scarabaeidae, Geotrupidae) communities across a 3–31 yearchronosequence of restored prairies, in which sites varied in the presence of re‐introduced bison and prescribed fire. We calculated functional diversity metrics and community‐weighted mean trait values using behavioural and morphological measurements. We also performed a dung decomposition experiment to measure an ecosystem function driven by these insects. 3. Bison presence doubled beetle abundance and increased richness by 50%. Shannon diversity increased with restoration age, nearly doubling from the youngest to oldest restorations. Functional diversity was unchanged by site characteristics, except functional richness, which was reduced by bison and fire presence. Beetles were, on average, smaller in older restorations, although this pattern was weaker when bison were present. 4. Dung decomposition was unaffected by site characteristics but increased with community weighted mean beetle mass. Dung decomposition was better predicted by mean trait values, suggesting that supporting large‐bodied species may be more important than species diversity in settings where maximizing decomposition function is a goal. 5. Restoration managers should consider dung beetle communities and their functional characteristics when making management decisions, particularly where large grazers are a component of management strategies.