Fig. 1

Jaccard-Tanimoto similarity distribution for each fingerprint across all possible pairwise comparisons in the natural product dataset. Violin plots indicate the percentiles of the distribution of Jaccard-Tanimoto similarities, with the circle indicating the median similarity value. The fingerprints where the similarity distribution on natural products is significantly different than the one obtained for drug-like compounds are highlighted in bold (Mann Whitney tests with Benjamini–Hochberg correction, α = 0.05)