Beyond microbes: mapping soil food web biodiversity via lipid fingerprinting

Unlocking the functional complexity of soil ecosystems requires robust methods for characterizing the diverse communities residing within them. While phospholipid fatty acid (PLFA) analysis has long served as a standard tool for quantifying soil microbial community composition, it is inherently constrained by low taxonomic resolution and a historical bias toward bacteria and fungi, often obscuring the contributions of other critical soil organisms. To address this gap, intact polar lipid analysis offers potentially deeper and more detailed insight into the soil (micro)biome. Here, we present a comprehensive lipidomic reference library built from intact polar lipids extracted from pure isolates and individual species spanning the tree of life, including bacteria, archaea, fungi, protists, plants, and soil fauna to enable high-resolution, culture-independent biodiversity assessments.

Using reversed phase UPLC separation followed by dual-polarity, high-resolution Orbitrap MS/MS and molecular networking, we captured over 140,000 molecular features and organized them into approximately 10,000 molecular families. This library covers ~30 phyla and >50 lipid classes, extending the analytical window far beyond PLFA to include diverse glycerophospholipids, glycolipids, sphingolipids, and neutral lipids. Hierarchical analyses reveal distinctive lipidomic architectures across taxonomic levels, with over half of the detected compounds appearing exclusive to single phyla. Beyond standard microbial signals, we identified rich lipid fingerprints specific for faunal and protist groups (e.g., Arthropoda, Nematoda, Mollusca, Amoebozoa), plastid-associated glycolipids in photosynthetic lineages, and characteristic archaeal membrane compositions. Clustering these features yielded thousands of phylum-exclusive molecular families, providing candidate biomarkers with built-in signal redundancy.

As a proof-of-concept, we detected these signatures in heterogeneous soil samples, supporting the feasibility of the approach while highlighting the need for broader validation of potential biomarker families. These findings establish a path toward high-resolution lipid-based mapping of soil community composition and food-web structure, offering a powerful, functional complement to existing genomic and biochemical approaches.