We first asked whether motif load would be able to detect the expected link between evolutionary and individual variation. We used a published metric, branch length score (BLS) [40], to characterize the evolutionary conservation of a motif instance. This metric utilizes both a PWM-based model of the conservation of bases and allows for motif movement. Reassuringly, mutational load correlated with BLS in both species, with evolutionarily non-conserved motifs (BLS = 0) showing by far the highest degree of variation in the population (Figure 3b). At the same time, approximately 40% of human and fly TFBSs with an appreciable load (L > 5e-3) still mapped to reasonably conserved sites (BLS > 0.2, approximately 50% percentile in both organisms), demonstrating that score-reducing mutations at evolutionarily preserved sequences can be tolerated in these populations.
