Functional-Taxonomic Scaling Resolves Conflicting Average Genome Size Estimates Across Environmental Gradients

Community-level average genome size (AGS) is a key trait linking microbial diversity, functional potential, and eco-evolutionary strategy, yet estimates of AGS diverge markedly among common methods. Here, we introduce and validate a framework that adjudicates conflicting AGS estimates by testing which estimate yields a gap between scaled functional diversity and scaled taxonomic diversity that is consistent with the expected positive association between genome size and gene functional breadth. Across independent environmental gradients, metagenomic AGS estimates align with this framework and expectations from metagenome-assembled genome (MAG) dynamics and gene co-occurrence network module sizes, whereas 16S rRNA metabarcoding-based AGS estimates deviate, particularly in extreme environments with sparse reference coverage. Applying this validated framework, we reveal a U-shaped relationship between soil pH and AGS across a broad pH range, unifying prior findings of decreasing AGS from acidic to neutral soils with a newly observed increase along an alkaline gradient. Applying the same framework to a saline-gradient dataset, we further suggest that conclusions of a recent study regarding eco-evolutionary trade-offs under salt stress may stem from method biases. This work identifies a potentially pervasive conflict in AGS estimation, provides a framework to resolve it, and clarifies and expands relationships between AGS and the environment.