Heavy metal (PTE) ecotoxicology, data review: Traditional vs. a compositional approach

Area of Science:

• Environmental Science
• Geochemistry
• Ecotoxicology

Background:

• Potentially Toxic Elements (PTEs), or heavy metals, are widespread in soils, posing environmental and health risks.
• Analyzing PTEs is complex due to soil interactions and their compositional nature, where data represent ratios summing to a whole.

Purpose of the Study:

• To evaluate the effectiveness of the centered log ratio (clr) transformation for analyzing PTEs.
• To compare clr with traditional log10 transformation and untransformed data in soil and animal tissue samples.

Main Methods:

• Applied clr, log10 transformations, and used untransformed data on Eurasian badger (Meles meles) tissues (liver, muscle, kidney) and soil data.
• Utilized cluster analysis, heatmaps, and principal component analysis (PCA) to assess PTE trends and relationships.

Main Results:

• clr and log10 transformations resolved compositional trends at the sample level, unlike untransformed data.
• Heatmaps revealed clr's ability to isolate PTE relationships and identify dataset commonalities, which log10 could not.
• PCA of clr data showed similarities in soft tissues and differences from soil, a distinction log10 failed to make.

Conclusions:

• The clr transformation consistently outperformed log10 and untransformed data across various analytical scenarios.
• A compositional approach using clr provides more realistic interpretations of PTEs in soil and animal tissues.