In addition the above work, we also re-implemented the aDNA simulator gargammel (which implements the Briggs model of aDNA damage) in a easier to use python package called pygargammel:

github.com/computations...

The main take away from the paper is that, of the sources of damage common to aDNA, the only one that really "matters" is fragment length. Other sources, such as single or double strand deamination, are not significant enough to strongly degrade placement accuracy.