Utrametricty precision decisions

[snacktavish]

https://github.com/OpenTreeOfLife/chronosynth/blob/main/chronosynth/chronogram.py#L76

A number of trees ages can't be calculated because they fail here with a tree ultrametricty error, under the default dendropy ultrametricty threshold (1e-05)
e.g. ot_1255@tree1

raise error.UltrametricityError(
dendropy.utility.error.UltrametricityError: Tree is not ultrametric within threshold of 1e-05: 0.00010000000000331966.
Encountered in subtree of node <Node object at 0x7f7e7016caf0: 'node10' (None)> (edge length of 181.521):

((Phrynus:73.8877,Damon:73.8877)node14:159.996,(Mastigoproctus:146.542,SchizomidaYucatan:146.542)node11:87.3416)node10:181.521

Age of children:

• <Node object at 0x7f7e7016cc70: 'node14' (None)>: has age of 73.8877 and edge length of 159.996, resulting in parent node age of 233.8837
• <Node object at 0x7f7e7016ccd0: 'node11' (None)>: has age of 146.542 and edge length of 87.3416, resulting in parent node age of 233.8836

However, can calc node ages if threshold is set to 0.001 or 0.,01

Any harm in this? What's 10,000 years among friends??

[hyanwong]

Big trees (lots of nodes from tip to root) are likely to compound errors (because of the newick format storing branch lengths rather than absolute times) more than smaller trees. So there's a reasonable justification for OpenTree, which you might assume would have very large trees sometimes, in having a larger threshold than the dendropy default.