Gene birth contributes to structural disorder encoded by overlapping genes

S. Willis; J. Masel

Note that frameshifted versions of high-ISD proteins have higher ISD than frameshifted versions of low-ISD proteins ( Figure 5A ). The two members of an overlapping pair share the same %GC content, and random sequences with higher %GC have substantially higher ISD [1] , so this could be responsible for the trait correlation. A facilitate-birth bias toward high ISD in newborn genes might do so in part via high %GC in the overlapping region at the time of birth, causing overlapping sequences to be biased not just toward high-ISD novel genes, but also toward high-ISD ancestral genes. Our 32% estimate attributes all of the ISD elevation in ancestral overlapping genes to constraint, but given the trait correlation shown in Figure 5A , some of this might also be due to birth facilitation, making 32% an underestimate. Note that novel genes born into the +2 frame have high ISD above and beyond the intrinsic correlation ( Figure 5B ), and are thus responsible for the statistical difference between ancestral vs. novel when frame is controlled for ( Figure 4B ).