7. To summarize, we see very different patterns of evolution in chronic infections, and suggest that only very rarely, does a chronic infection lead to a transmissible new variant.

6. In fact, one other patient showed an ns rate of zero. In fact, in this patient we saw many point deletions and insertions. These should in theory represent defective (“dead”) virus. We think that it is possible that defective virus thrives due to co-infections that allow rescuing such viruses.

5. However, the rate of non-synonymous (ns) divergence varied widely among patients. In fact, there was only one patient where the ns rate was a lot higher than the syn rate, which is an indication for dramatic adaptive evolution.

4. When focusing on synonymous (syn) mutation frequencies, we showed that the rate of syn divergence is ~2X1E-06 mutations/base/day, consistent with previous estimates from acute infections and from tissue culture.

3. We followed mutation frequencies in a cohort of nine patients, all of whom were immunocompromised and experienced prolonged SC2 infections.

2. We showed (not for the first time) that the sequencing process itself introduces errors, and this is especially pronounced in low viral load samples. We thus sequenced all samples in duplicate, allowing us to remove sequencing errors.

1. Why are chronic SC2 infections interesting? Mainly, since it is thought that variants-of-concern (VOCs) derived from such infections, since they display high rates of evolution and similar patterns of mutations. But how often do chronic infections lead to highly divergent variants?