Answers are grounded in this paper and cite their sources.
Soil microorganisms mediate nutrient turnover, carbon storage, disease suppression, and plant stress tolerance, yet the evidentiary distance between microbial measurements and agronomic outcomes remains poorly defined. An evidence map is needed that states exactly what has been reported, in which study settings, and with what reporting gaps. An exploratory evidence map of PubMed-indexed records published from 1 January 2015 to 29 September 2026 was conducted. A Boolean query combining soil, rhizosphere, and plant microbiome concepts with crop and agricultural terms was executed on 29 September 2026 through the NCBI E-utilities. Records were deduplicated by identifier and normalised title, screened with prespecified automated title/abstract rules, and classified with lexical tags for evidence domain, study design, geography, intervention class, and crop type. Tag accuracy was quantified on 194 manually annotated records drawn from seven strata, and precision, recall, and 95% Wilson confidence intervals were computed. Of 11,061 records identified, 4 duplicates and 573 records excluded at screening left 10,484 screen-positive records. Tags indicated microbial inoculants (6,695; 63.9%), climate or environmental stress (2,748; 26.2%), nutrient cycling (2,453; 23.4%), and plant health or disease (2,377; 22.7%). An explicit field design was stated in 1,256 records (12.0%), with a validation-adjusted field share of 22.1% (95% CI 13.1%—37.5%). Economic outcomes appeared in 593 records (5.7%) and negative or null result signals in 392 (3.7%). Manual validation gave precision of 93.3% for eligibility, 90.0% for the inoculant tag, 84.0% for the stress tag, and 100% for geographic tagging. Within the PubMed-indexed literature, field verification, economic outcomes, and negative or null results are under-represented relative to controlled-environment and molecular studies. The map defines a PubMed-specific retrieval landscape, and the quantified gaps specify what a subsequent multi-database, study-level systematic review should verify.
No takes yet. Share an insight, caveat, or question.
Mamun et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: