I thought to examine signal of natural evolution in #nCoV2019 by looking at distribution of mutations in the lineage leading to nCoV and compare this distribution to mutations occurring in other SARS-like bat viruses 1/9
I used the evolutionary tree here nextstrain.org/groups/blab/sa… and the bioinformatic pipeline here github.com/blab/sars-like… to identify mutations leading to #nCoV2019 and also identify mutations leading to bat virus RaTG13, which is its closest sequenced relative. 2/9
The nCoV lineage is shown as a dashed line under the "nCoV" label (lower right). The RaTG13 lineage is shown as a dashed line above the "RaTG13" label (also lower right). 3/9 
These two viruses are separated by ~1100 nucleotide changes, consistent with 20-70 years of evolution between them (assuming an evolutionary rate of between 3x10^-4 and 1x10^-3 subs per site per year). This suggests a common ancestor in bats between 1950 and 2000. 4/9
The #nCoV2019 lineage has 79 amino acid changes and 554 nucleotide changes relative to this common ancestor, so that only 14.3% of changes result in differences in amino acid. 5/9
Evolutionary biologists commonly use the ratio of amino acid changes to nucleotide changes to look for natural selection (en.wikipedia.org/wiki/Ka/Ks_rat…). Observing 14% of changes as amino acid changes fits with common expectations for sequences evolving under purifying selection. 6/9
And in fact we see that the RaTG13 lineage has 79 amino acid changes and 563 nucleotide changes for a ratio of 14.2%. Thus we see that the outbreak virus has a similar distribution of nucleotide vs amino acid changes as its closest bat virus relative. 7/9
Additionally, I thought to look at the distribution of amino acid changes across genes in the virus genome. In this case, we see highly similar distributions between the nCoV lineage and its closest bat virus relative. 8/9
Looking at mutation distributions, it appears that the genetic differences in #nCoV2019 are consistent with differences expected to arise during natural evolution, as I would expect engineering to have a distorted AA to nuc ratio and to focus on changing a subset of genes 9/9
If you'd like to investigate yourself, you can either run the entire openly available bioinformatics pipeline here github.com/blab/sars-like… or just work from the resulting JSON file that contains all of these mutation calls here raw.githubusercontent.com/blab/sars-like…
