Ebola dataset broken

[jameshadfield]

Possible causes:

• tree time confidence intervals mis-parsed in Auspice
• globals are specific to zika

[rneher]

not sure what exactly is the issue, but I suspect the you are referring o confidence intervals not overlapping the actual tree node. this can occur in data sets like ebola with very many zero-length branches. in such a case, the most-likely configuration is a very atypical one. (example: the 5-95% confidence of a draw from an exponential distribution does not contain the most likely values). This should be fixed by assigning the marginal ml positions rather than the global ml positions.

[jameshadfield]

Thanks @rneher
I'll look into this ASAP (I think there are more issues than just the CIs)

Current appearance of the ebola tree:

Timelines of dataset change (c. 6s):

[rneher]

set this to false:

https://github.com/nextstrain/auspice/blob/master/src/components/tree/treeView.js#L194

the confidence stuff was just a first shot, prob shouldn't have made its way into master yet.

[jameshadfield]

Closed in PR #156