Figure S1: Sampling strategy along the Centralia temperature gradient
Twelve surface soils were collected along two fire fronts. Sampling sites are classified based on historical fire activity (Elick 2011) and observations of fire activity at the time of sampling: fire affected (red), recovered (yellow), and recovered (green, reference). Red bullseye indicates fire origin, and fire fronts one and two are indicated with arrows (F1 and F2, respectively).
Figure S2: Comparison of community structure assessed using two different methods
Community structure determined by rplB (A) is similar to previously described community structure determined by 16S rRNA gene sequencing reported in Lee and Sorensen et al. 2017(B). Samples are classified by their fire history: fire affected (n = 6), recovered (n = 5), and reference (n = 1).
Figure S3: Pair-wise Spearman’s correlations of normalized ARG abundances in Centralia
Spearman’s rho is indicated in each cell and by color, where negative correlations are red and positive correlations are blue. False discovery rate adjusted significance is noted by asterisks.
Figure S4: Relationship between normalized abundance of ARGs and soil temperature
Point shape indicates soil fire classification. Coverage-adjusted abundance for each gene was normalized to total abundance of single copy gene rplB. Normalized abundance is plotted against soil temperature. Note the differences in y-axes. Shape indicates soil classification based on fire history.
Figure S5: Beta diversity of Centralia microbial communities with rplB and ARGs
Principal coordinate analysis (PCoA) based on weighted Bray-Curtis distances of community structure (A) and ARG structure (B). Colors represent soil temperature, and shape indicates soil classification based on fire history.