Beyond Taxonomic Identification: Integration of Ecological Responses to a Soil Bacterial 16S rRNA Gene Database

Abstract

High-throughput sequencing 16S rRNA gene surveys have enabled new insights into the diversity of soil bacteria, and furthered understanding of the ecological drivers of abundances across landscapes. However, current analytical approaches are of limited use in formalizing syntheses of the ecological attributes of taxa discovered, because derived taxonomic units are typically unique to individual studies and sequence identification databases only characterize taxonomy. To address this, we used sequences obtained from a large nationwide soil survey (GB Countryside Survey, henceforth CS) to create a comprehensive soil specific 16S reference database, with coupled ecological information derived from survey metadata. Specifically, we modeled taxon responses to soil pH at the OTU level using hierarchical logistic regression (HOF) models, to provide information on both the shape of landscape scale pH-abundance responses, and pH optima (pH at which OTU abundance is maximal). We identify that most of the soil OTUs examined exhibited a non-flat relationship with soil pH. Further, the pH optima could not be generalized by broad taxonomy, highlighting the need for tools and databases synthesizing ecological traits at finer taxonomic resolution. We further demonstrate the utility of the database by testing against geographically dispersed query 16S datasets; evaluating efficacy by quantifying matches, and accuracy in predicting pH responses of query sequences from a separate large soil survey. We found that the CS database provided good coverage of dominant taxa; and that the taxa indicating soil pH in a query dataset corresponded with the pH classifications of top matches in the CS database. Furthermore we were able to predict query dataset community structure, using predicted abundances of dominant taxa based on query soil pH data and the HOF models of matched CS database taxa. The database with associated HOF model outputs is released as an online portal for querying single sequences of interest (https://shiny-apps.ceh.ac.uk/ID-TaxER/), and flat files are made available for use in bioinformatic pipelines. The further development of advanced informatics infrastructures incorporating modeled ecological attributes along with new functional genomic information will likely facilitate large scale exploration and prediction of soil microbial functional biodiversity under current and future environmental change scenarios.

Building a genome-based understanding of bacterial pH preferences

“…S3). Our results are in line with previous work that found no phylogenetic conservation of pH preferences across both cultured bacteria ( 14 ) and uncultured bacteria ( 40 ). Neither taxonomic nor phylogenetic information are generally useful for determining pH preferences without additional information, highlighting the potential value of incorporating functional gene information to better predict bacterial pH preferences.…”

“…S7). Note that the estimated pH preferences in that study were obtained using a different approach from our study, and that the pH preferences for many taxa were inferred using models with relatively weak fits, a point the authors were careful to acknowledge ( 40 ). Despite the limitations of this independent dataset and the limitations associated with our model, the correspondence between predicted and observed pH preferences from this independent study (and an independent set of genomes from our study, Fig.…”

“…We further validated our model using information on bacterial pH preferences estimated in a completely independent study of bacterial distributions in soils across the UK ( 40 ). The linear regression between the statistically estimated pH preference of taxa in the dataset and our model predictions had R 2 = 0.21 and MAE = 0.93 (Fig.…”

“…We further validated our model using information on bacterial pH preferences estimated in a completely independent study of bacterial distributions in soils across the UK (40).…”

The Evolution of Ecological Diversity in Acidobacteria

“…So far, the HOF modeling approach (Jansen and Oksanen, 2013), which uses hierarchical logistic regressions to find niche optima along environmental gradients, has primarily been applied to eukaryotes. Only recently this approach has successfully also been applied to soil bacteria (Tripathi et al, 2018;Jones et al, 2021), although soil heterogeneity at the microscale level (micrometer) cannot be taken into account. To test our approach, we compared the optimum niche value predicted for pH for individual OTUs with the corresponding pH growth optimum determined in laboratory cultures of isolates from the same OTU.…”