Fig 1.
Collector’s curve and proportion analysis of species or enzyme prevalence.
(A) Number of observed species (orange points) or enzymes (green points) as sample number increases. (B) Histogram of the proportion of species (orange bars) or enzymes (green bars) present in each sample, binned by the proportion of samples. Values above each bar are the number of species or enzymes in that bin.
Table 1.
Summary statistics of XGBoost modeling results.
Data included is from all 25 models and each soil health indicator. RMSE = root mean squared error. No. of Enzymes represents the number of enzymes retained in the model after feature selection.
Fig 2.
Enzymes with the top ten average gains for each soil health indicator rating.
For each enzyme on the x-axis, the average gain on the y-axis was calculated by averaging the gains from all 25 XGBoost models for the respective indicator (see section Molecular index development and testing for method of gain calculation). Enzymes are labelled by their Enzyme Commission numbers and additional details about each enzyme can be found in S1 and S2 Tables.
Fig 3.
Number of enzymes associated with each KEGG pathway for each soil health indicator rating.
Each enzyme may be classified to more than one pathway and the reported values are expressed as a proportion of the total classifications. The “Other” pathway represents pathways that were irrelevant to enzyme function in soils (i.e., pathways that were human-related).
Fig 4.
MISH indicator ratings between SEMWISE indicator bins for each soil health indicator.
SEMWISE Indicator Bins represent the individual indicator score in which “very low” is the lowest soil health and “very high” is the highest soil health based on the SEMWISE model described in Deel et al. [34]. The MISH Indicator Ratings were calculated using the top enzymes for each indicator (see section Molecular index development and testing for calculation) and are expressed as a percentage between 0–100. Different letter labels indicate a significant difference between bins.
Fig 5.
MISH Overall Rating vs SEMWISE and SHMI rating bins.
Each panel shows the distribution of MISH Overall Ratings binned by (A) soil health (SEMWISE) rating or (B) management (SHMI rating). Both ratings on the x-axes are described in Deel et al. [35]. Briefly, the SEMWISE Rating represents the soil health score that incorporates all soil health indicators. The SHMI rating represents the overall management score based on the soil health principles. In both ratings, the “very low” bin represents the lowest soil health and “very high” bin represents the highest soil health. The MISH Overall Rating was calculated using top enzymes from all indicators (see section Molecular index development and testing for calculation) and is expressed as a percentage between 0–100. Different letter labels indicate a significant difference between bins. Points are color-coded by current land use with annual cropland (red), perennial cropland (green), and rangeland (blue).