Figure 4.

The frequency of trimers in the sequence context near errors. Each trimer is the three reference bases leading up to, and including, an error site. As described in the ‘Materials and Methods’ section, we counted the occurrences of each trimer, then normalized by the abundance of the corresponding trimer in the genome. A normalized frequency of 1 would mean the trimer is associated with errors exactly as much as in the null hypothesis where errors are randomly distributed. Trimers are presented in order of their median frequency among all platforms.