Extended Data Fig. 2: Comparison of Gnocchi score between coding and non-coding regions.
From: A genomic mutational constraint map using variation in 76,156 human genomes
a, The proportion of highly constrained windows (Gnocchi ≥ 4) as a function of the percentage of coding sequences in a window (left to right: N = 1,906/49,525, 3,244/55,676, 2,240/18,461, 1,506/7,094, 969/3,519, 569/1,946, 364/1,223, 283/910, 243/724, 10,392/30,138). The intervals (x-axis) are left exclusive and right inclusive. “Exonic only” refers to the 1kb windows created from directly concatenating coding exons into 1kb sequences. Error bars indicate standard errors of the proportions. b, The exonic-only regions (N = 27,875; purple) present a significantly higher Gnocchi score than regions that are exclusively non-coding (N = 1,843,559; blue). Dashed lines indicate the medians. c, The proportion of highly constrained windows (Gnocchi≥4) as a function of the proportion of exonic windows being added to the dataset of non-coding windows. d, Gnocchi score percentiles of non-coding versus exonic windows. About 0.05% (100-99.95%) and 3.12% (100-96.88%) of the non-coding windows exhibit similar constraint to the 90th and 50th of exonic regions, respectively.
