# Contribution of Range‐Wide and Short‐Scale Chemical Soil Variation to Local Adaptation in a Tropical Montane Forest Tree

## TL;DR

This study shows how soil chemistry influences the genetic adaptation of a tropical tree species across different spatial scales.

## Contribution

The paper reveals that soil chemical variation drives local adaptation in trees through distinct genetic mechanisms at range-wide and local scales.

## Key findings

- 49 and 23 candidate SNPs were identified at the range-wide and local scales, respectively, with little overlap.
- Polygenic models explained ~20% of variation in soil properties like Ca2+, EC, and pH at different scales.
- Geography and population isolation were key in explaining genetic-soil co-variation.

## Abstract

Local adaptation is a fundamental process that allows populations to thrive in their native environment, often increasing genetic differentiation with neighboring stands. However, detecting the molecular basis and selective factors responsible for local adaptation remains a challenge, particularly in sessile, non‐model species with long life cycles, such as forest trees. Local adaptation in trees is not only modeled by climatic factors, but also by soil variation. Such variation depends on dynamic geological and ecological processes that generate a highly heterogeneous selective mosaic that may differentially condition tree adaptation both at the range‐wide and local scales. This could be particularly manifest in species inhabiting mountain ranges that were formed by diverse geological events, like sacred fir (Abies religiosa), a conifer endemic to the mountains of central Mexico. Here, we used landscape genomics approaches to investigate how chemical edaphic variation influences the genetic structure of this species at the range‐wide and local scales. After controlling for neutral genetic structure, we performed genotype‐environment associations and identified 49 and 23 candidate SNPs at the range‐wide and local scales, respectively, with little overlap between scales. We then developed polygenic models with such candidates, which accounted for ~20% of the range‐wide variation in soil Ca
2+ concentration, electric conductivity (EC), and pH, and for the local variation in soil EC and organic carbon content (OC). Spatial Principal Component Analyses further highlighted the role of geography and population isolation in explaining this genetic‐soil co‐variation. Our findings reveal that local adaptation in trees is the result of an intricate interaction between soil chemical properties and the local population's genetic makeup, and that the selective factors driving such adaptation greatly vary and are not necessarily predictable across spatial scales. These results highlight the need to consider edaphic variation in forest genetic studies (including common garden experiments) and in conservation, management and assisted migration programs.