raxml returns higher likelihood for constrained tree than unconstrained tree for some simulated matrices
Closed this issue · 2 comments
caseywdunn commented
This is a known issue, and is addressed here:
https://groups.google.com/forum/#!topic/raxml/qn7_ZXoJTHg
We should redo some of hte problematic searches with the addition of the --no-bfgs
raxml option. If that fixes the problem for those matrices, we should rerun the tests in the manuscript that were impacted by this. If they work, we should revise the manuscript to reflect the fix.
shchurch commented
Currently testing if removing BFGS protocol resolves the issue. Preliminary results indicate that yes. See https://groups.google.com/forum/#!topic/raxml/qn7_ZXoJTHg
caseywdunn commented
RAxML restored to default lieklihood engine.