aDNA damage comments

[mattiassherman]

Hi, really cool website! I just wanted a give a couple comments on the aDNA damage section of the population genomics tab so that it can be as accurate as possible.

First, concerning fragmentation, perhaps a little detail but it results from the breaking of chemical bonds between the base and the sugar-phosphate backbone and then of the backbone to produce a gap, not from broken chemical bonds between bases (Dabney et al 2013). This process (depurination+beta-elimination) results in gaps (also called single-stranded breaks) and not necessarily double-strand breaks. However, multiple of these events on both strands will result in fragmentation (and they may well result in double-strand breaks, but more often in overhangs). Additionally, nicks do not refer to missing bases (these are abasic sites/AP sites), but to a break in the sugar-phosphate backbone (without base loss). I don't think this is well understood in aDNA, and hasn't even been formally proven to occur (Bokelmann et al 2020 do mention them).

Second, the explanation for the cytosine deamination mismatch patterns is somewhat misleading, but that is probably because the literature on it is quite confusing. In single-stranded libraries, during strand fill-in, uracil-tolerant polymerases incorporate adenines across from uracils, resulting in C-to-T mismatches (the reverse strand bases are inversed). In double-stranded libraries, during end-polishing, U-tolerant polymerases will fill-in 5' overhangs, leading to the incorporation of adenines across from uracils (but only at these 1 to 2 nucleotide-long overhangs). During adapter fill-in, strand-displacing and U-tolerant polymerases will extend from any nick or gap in the double-stranded DNA molecule, resulting in the incorporation of adenines across from uracils and the removal of C-to-U mismatches due to the strand displacement (Briggs et al 2007). It isn't clear why this fill-in process happens more towards the 3' ends of DNA strands, but I would guess that it's likely due to the polymerases more often starting this fill-in at strand ends (there is some evidence of depurination occurring more often closer to strand ends, An et al 2014) and more quickly at the 3' ends (just because polymerases work on the 5' to 3'). Then during the first replication of DNA templates with a U-tolerant polymerase, uracils will be read as thymines and adenines incorporated across from uracils during end-polishing adapter fill-in will result in synthetic misincorporations reflecting 5' C-to-U mismatches.

And lastly, I would suggests adding "if not accounted for" after "error" in the last sentence of the DNA damage section, just to hint at the different possible ways of dealing with damage (experimentally or bioinformatically).

Hope this is helpful! My goal is to reduce the confusion around aDNA damage in the field, so hopefully this can help you make your impressively detailed website as accurate as it can be! Do let me know if any of this is unclear.

Mattias