Assertion `tip_index == tip_count' failed

Question

Assertion `tip_index == tip_count' failed

Opened this issue a month ago · 8 comments

Hello,

I've encountered another issue while running the tool. I got the following error message:

/usr/bin/bash: line 22: 46055 Aborted (core dumped) raxml-ng --redo --msa results/fasta/family/45-of-79/aligned.fa --outgroup $OG --model GTR+G --tree-constraint results/fasta/family/45-of-79/remapped.tre --workers 1 --blopt nr_safe --search > logs/run_raxml/run_raxml-45-of-79.log 2>&1
raxml-ng: /opt/conda/conda-bld/raxml-ng_1714512705109/work/libs/pll-modules/libs/libpll/src/parse_utree.y:508: utree_wraptree: Assertion `tip_index == tip_count' failed.

Do you know what it means? I've attached the relevant files and logs.
45-of-79.zip

Thanks,
Dominika

Answer 1 · 2024-11-19T10:34:03.000Z

Hi Dominika,

could you please also attach logs/run_raxml/run_raxml-45-of-79.log file?

Thanks!

Answer 2 · 2024-11-19T10:40:49.000Z

Hello,

Here is the log file
run_raxml-45-of-79.log

Answer 3 · 2024-11-19T10:51:21.000Z

sorry, but it seems to be from a different run: it did not use constraint and finished without errors,

Answer 4 · 2024-11-19T12:09:02.000Z

it is the same run, I don't know why the errors are not in the log file, it is shown in the terminal, that's why I pasted it separately

edit: the log file I attached in the previous comment is indeed from a different run, the original log file is in the zipped file (aligned.fa.raxml), the ".log" extension has been cut for some reason

Answer 5 · 2024-11-19T13:30:24.000Z

I do not see aligned.fa.raxml file, only aligned.fa.raxml.log. And with the command therein I can't reproduce the error above, BUT I get segfault as in your previous bug report. And it is for the same reason: your constraint tree does only have 3 taxa, which is not supported (but sure, there must be a check and a clear error message in such cases).

In any case, I assume the purpose of your constraint is to enforce the monophyly of 3 taxa in the outgoup (AADBE118-10,AADBE232-10,AAI123-13), right?

If so, your constraint tree must look like this:

((AADBE232-10,AAI123-13),AADBE118-10,(AADBE243-10,AADBE214-10,AAPY004-10,AADBE123-10,AADBE307-10,AADBE173-10,AADBE242-10,AADBE121-10,AADBE252-10,AADBE256-10,AADBE258-10,AADBE259-10,AADBE135-10,AADBE213-10,AADBE265-10));

i.e. it has to include all taxa.

Although raxml-ng does support non-comprehensive constraint trees, they have a different meaning: all taxa not present in the constraint tree are considered "free", i.e. they can be placed anywhere. also inside the constraint tree clades.

Hope this helps!

Answer 6 · 2024-11-21T17:56:03.000Z

Hi Dominika,

it would be great if you could try to reproduce the problem with the latest development branch:
https://github.com/amkozlov/raxml-ng/wiki/Installation#building-development-branch

Thank you :)

Answer 7 · 2024-11-28T09:41:27.000Z

Hello!

Thanks for the update and help, I will rerun it shortly!
I have one more question, what is the most efficient way to execute the tool in terms of splitting resources? Would it be better to run 10 processes simultaneously with 12 threads each, or more processes with 1 thread each? Or it doesn't make a big difference? Right now it takes ages to execute the pipeline so I'm looking for a way to enhance the speed since my input dataset is large.

Thank you

Answer 8 · 2024-11-29T11:14:58.000Z

Hi Dominika,

please have a look at: https://github.com/amkozlov/raxml-ng/wiki/Parallelization

In most cases, automatic parallelization does a good job predicting the most efficient configuration. So you can just use --threads auto{N}, where N is the number of physical CPU cores on your system.

Also, the dev version I linked above should be 2x-5x faster thanks to using adaptive search heuristic described here: https://academic.oup.com/mbe/article/40/10/msad227/7296053?login=false