Can We Estimate Functionality of Soil Microbial Communities from Structure-Derived Predictions? A Reality Test in Agricultural Soils
Claudia Breitkreuz,Anna Heintz-Buschart,François Buscot,Sara Fareed Mohamed Wahdan,Mika Tarkka,Thomas Reitz,Thomas ReitzaDepartment of Soil Ecology,UFZ—Helmholtz Centre for Environmental Research,Halle,GermanybGerman Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig,Leipzig,Germany,Jeffrey A. Gralnick
DOI: https://doi.org/10.1128/Spectrum.00278-21
IF: 3.7
2021-08-05
Microbiology Spectrum
Abstract:Computational approaches that link bacterial 16S rRNA gene amplicon data to functional genes based on prokaryotic reference genomes have emerged. This study aims to validate or refute the applicability of the functional gene prediction tools for assessment and comparison of community functionality among experimental treatments, inducing either fast or slow responses in rhizosphere microbial community composition and function. Rhizosphere samples of wheat and barley were collected in two consecutive years at active and mature growth phases from organic and conventional farming plots with ambient or future-climate treatments of the Global Change Experimental Facility. Bacterial community composition was determined by 16S rRNA gene amplicon sequencing, and the activities of five extracellular enzymes involved in carbon (β-glucosidases, cellobiohydrolase, and xylosidase), nitrogen ( N -acetylglucosaminidase), and phosphorus (acid phosphatase) cycles were determined. Structural community data were used to predict functional patterns of the rhizosphere communities using Tax4Fun and PanFP. Subsequently, the predictions were compared with the measured activities. Despite the fact that different treatments mainly drove either community composition (plant growth phase) or measured enzyme activities (farming system), the predictions mirrored patterns in the treatments in a qualitative but not quantitative way. Most of the discrepancies between measured and predicted values resulted from plant growth stages (fast community response), followed by farming management and climate (slower community response). Thus, our results suggest the applicability of the prediction tools for comparative investigations of soil community functionality in less-dynamic environmental systems. IMPORTANCE Linking soil microbial community structure to its functionality, which is important for maintaining health and services of an ecosystem, is still challenging. Besides great advances in structural community analysis, functional equivalents, such as metagenomics and metatranscriptomics, are still time and cost intensive. Recent computational approaches (Tax4Fun and PanFP) aim to predict functions from structural community data based on reference genomes. Although the usability of these tools has been confirmed with metagenomic data, a comparison between predicted and measured functions is so far missing. Thus, this study comprises an expansive reality test on the performance of these tools under different environmental conditions, including relevant global change factors (land use and climate). The work provides a valuable validation of the applicability of the prediction tools for comparison of soil community functions across different sufficiently established soil ecosystems and suggest their usability to unravel the broad spectrum of functions provided by a given community structure.
microbiology