Absrel error

[WRran-Hydrogen]

Dear @spond ,

There is error in absrel model in "All branches to test", when running to the last step. The following are running screenshots details: The log file was overwritten by running it again. Can you help me solve this problem? Many appreciate you!

[spond]

Dear @WRran-Hydrogen,

Try adding ENV="TOLERATE_NUMERICAL_ERRORS=1;" as a command line argument. Can you also confirm what HyPhy version you are using (hyphy --version)?

Best, Sergei

[WRran-Hydrogen]

Dear @spond,

The version detail is HYPHY 2.5.2(MP) for Linux on x86_64 and I try to add ENV="TOLERATE_NUMERICAL_ERRORS=1;" as a argument. But since I didn't set the branch label, absrel will test all the branches and take a lot of time, so it might be 12h before the results come out. plus can I ask a navie question? Does my alignment sequence need to be trimmed based on the quality of the gap? Is there any good way to do this that would significantly reduce the amount of computation and improve the quality of the codon aligned sequence? I am currently using the program macse reportMaskAA2NT to trim the aligned codon sequences using the protein aligned trimmed sequences. Thanks for your reply!

Cheers!

[spond]

Dear @WRran-Hydrogen,

2.5.2 is ~5 years old. I strongly encourage you to update to a newer HyPhy version. I am also pretty sure that for something as old as 2.5.2 the solution I proposed won't work; newer versions are also much faster.

Yes, you should definitely try to improve alignment quality. There's no gold standard here; my recent suggestions are here: https://hackmd.io/g387lOT1SmudOueIeKhkuA

Best, Sergei

[WRran-Hydrogen]

Dear @spond,

Thanks for your help. I have updated Hyphy and am going to try again. But I entered the website Hackmd you provided and could not find the relevant process (maybe I need to follow to see it?). In addtion, Is it a better choice to select monophyletic branches for analysis of gene trees that contain great diversity (Excessive phylogenetic diversity seems to lead to a bias in positive selection due to an excessive number of aligned sequences)? Many thanks again!

Cheers.

[WRran-Hydrogen]

Dear @spond, I reran the Absrel model and it still produces the error, see the screenshot below:

command: hyphy absrel --alignment ARG_domain.msa --tree maxcc.contree --output out.json ENV="TOLERATE_NUMERICAL_ERRORS=1"

error.log:

Process file:

My hyphy version is HYPHY 2.5.29(MP) for Linux on x86_64, I don't know what the problem is, could you please help me?

[spond]

Dear @WRran-Hydrogen,

1. I would recommend updating to 2.5.59 (2.5.29 is now > 3 years old).
2. Add ENV="TOLERATE_NUMERICAL_ERRORS=1;" to the command line arguments to treat the errors you report as warnings.

Best, Sergei