Fig. 5

The distributions of the signature distances (LHS) \(d_{\text {sn}}(s,s'),d_{\text {sn}}(s,s'')\) and the base-pair distances (RHS) \(d_{\text {bp}}(s,s'),d_{\text {bp}}(s,s'')\) obtained from modular splitting (blue) and random splitting (orange). We generated 8000 random sequences \({\mathbf {x}}\) of length 500, and computed their structures s (MFE). For any two positions i and j, we compute \(s'\) (modular), \(s''\) (random) via ViennaRNA [12]. The red dashed line (left) denotes “threshold distance”, 31 (see main text)