a, Heat map of the top-15 genes enriched in the palaeofaeces, industrial and non-industrial samples (complete results in Supplementary Table 8). Functions were annotated using PROKKA38 (one-tailed Wilcoxon rank-sum tests with Bonferroni correction). The reads per kilobase per million reads (RPKM) values shown are on a log scale and scaled by row. An unscaled heat map is shown in Extended Data Fig. 12. b, Volcano plots showing enriched CAZymes signatures (two-tailed Wilcoxon rank-sum test with FDR correction) comparing palaeofaeces and non-industrial samples (left), palaeofaeces and industrial samples (middle), and non-industrial and industrial samples (right). Each data point represents a CAZy family. CAZymes are colour-coded according to manually annotated broad substrate categories. The horizontal dashed red line indicates adjusted P = 0.05. The vertical dashed red line indicates log2-transformed fold change = 0. For the left and middle plots, both the entire dataset and a magnified version are shown. For the right plot, the x-axis limits were set to −5 and 5 (as a result, eight statistically non-significant CAZymes were removed).