Fig. 2

Teleost intron size predicts mammalian intron size. A, median log₂ intron sizes in teleosts (x-axis) plotted against those in mammals (y-axis) indicate a strong correlation in size for longer introns. The vast majority of introns in both teleosts and mammals are longer than 75 bases (indicated by the dashed red lines); however we observe indicated sizes shorter than these in both clades that are likely to represent annotation artefacts rather than real short introns. B, and C, two-dimensional histograms for points in A. Colours range from dark blue (low) to bright pink (high) and indicate either the log transformed counts of the number of points within the indicated ranges (B) or, counts scaled by column to have means and standard deviations of 0 and 1 respectively. D, quantile plot of log₂ median teleost and mammal intron sizes. 20% of introns have a median length of 10² (1024) bp or more in teleosts. E, proportion of introns having a median length above the indicated quantiles in mammals plotted against log₂ teleost median length. Of the 20% of introns whose median teleost length are longer than 2¹⁰ bp, 73 and 20% have mammal lengths longer than 50 and 95% of mammalian introns (see right inset). F, excess of long introns in mammals (log₂ observed / expected ratios) for sets indicated in E