Tissue-Specific Enhancer Mutability Analysis

The size of enhancer regions differs widely between tissues. Furthermore, the mutability of the tissue-specific enhancer region differs significantly. The mutability of FBSEs, human gain enhancers, and adult brain-specific enhancers was estimated using the previously defined framework for DNMs (34 (link)). The framework for the null mutation model is based on the tri-nucleotide context, where the second base is mutated. Using this framework, the probability of mutation for each enhancer was estimated based on the DNA sequence of the enhancer. The probability of mutation of all the enhancers within the enhancer set (FBSEs, human gain enhancers, and adult brain subsections) was summed to estimate the probability of mutation for the entire enhancer set. The sequence composition and overall size in base pair vary significantly between fetal brain enhancers, human gain enhancers, and enhancers from adult brain subsections; hence, they may have different background mutation rates. To perform a valid comparison between the observed number of DNMs between fetal brain enhancers and adult brain enhancers, we normalised to the background mutation rate of fetal brain enhancers.
For example, the background mutation rate for FBSEs is 0.970718 and we observed 53 DNMs. Similarly, the background mutation rate for the adult brain subsection angular gyrus is 0.680226 and we observed 22 DNMs. Because of the difference in background mutation rate, we cannot directly compare the number of DNMs between fetal brain enhancers and angular gyrus. Hence, we normalised the observed number of DNMs in angular gyrus enhancers to a background mutation rate of 0.970718 using the following formula. $$\frac{\left(\text{Observed} \text{number} \text{of} \text{DNMs} \text{in} \text{angular} \text{gyrus} \text{enhancers}\times \text{mutation} \text{rate} \text{of} \text{fetal} \text{brain} \text{enhancers}\right)}{\text{mutation} \text{rate} \text{of} \text{angular} \text{gyrus} \text{enhancers}},$$ $\frac{\left(22\times 0.970718\right)}{0.680226} =31.395$
Similarly, we normalised the observed number of mutations from all adult brain subsections and human gain enhancers to a background mutation rate of 0.970718 (Table S6) so that we could perform a valid comparison between the observed number of mutations from various enhancer sets. The significance level between DNMs observed in FBSEs (n = 2) and adult brain enhancers (n = 8) was calculated using a two-tailed t test.