Scholarlink

<p>Soil microbial communities are central to ecosystem functioning, yet their large-scale diversity patterns and environmental sensitivities remain poorly understood. Here, we analysed a national soil microbiome dataset spanning all major Australian ecosystems to assess bacterial and fungal richness across diverse environmental gradients. Using supervised deep autoencoders—neural networks that compress complex data to identify key patterns—and structural equation models, we explained around 60% of the variance in richness. Bacteria and fungi showed contrasting spatial patterns, fungal richness was tightly constrained by moisture and organic carbon, while bacterial richness peaked in nitrogen-rich, topographically complex regions and persisted across broader conditions. We propose the bacterial-to-fungal richness ratio as a spatially explicit, scalable indicator of soil community composition and ecological condition. This ratio captures gradients of aridity, nutrient imbalance, and land-use intensification, and may help identify ecosystems vulnerable to environmental change, bridging microbial biogeography with applied soil health assessment across global biomes.</p><p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Decoding bacterial and fungal richness with autoencoders yields a unified ratio indicating soil health and ecological susceptibility

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月19日

查看全文
  • Raphael A. Viscarra Rossel,
  • Thorsten Behrens,
  • Andrew Bissett,
  • Lewis Walden,
  • Mingxi Zhang

摘要

Soil microbial communities are central to ecosystem functioning, yet their large-scale diversity patterns and environmental sensitivities remain poorly understood. Here, we analysed a national soil microbiome dataset spanning all major Australian ecosystems to assess bacterial and fungal richness across diverse environmental gradients. Using supervised deep autoencoders—neural networks that compress complex data to identify key patterns—and structural equation models, we explained around 60% of the variance in richness. Bacteria and fungi showed contrasting spatial patterns, fungal richness was tightly constrained by moisture and organic carbon, while bacterial richness peaked in nitrogen-rich, topographically complex regions and persisted across broader conditions. We propose the bacterial-to-fungal richness ratio as a spatially explicit, scalable indicator of soil community composition and ecological condition. This ratio captures gradients of aridity, nutrient imbalance, and land-use intensification, and may help identify ecosystems vulnerable to environmental change, bridging microbial biogeography with applied soil health assessment across global biomes.