search

FePhyFoFum / phyx

phylogenetics tools for linux (and other mostly posix compliant) computers
blackrim.org
GNU General Public License v3.0
111 stars 17 forks source link

strange behavior with pxbp on computing cluster #174

Open kateuckele opened 1 year ago

kateuckele commented 1 year ago

Pxbp version 1.3. I believe phyx was binary compiled, but I was not the person who did this. All the phyx commands that I have used up until now have not produced any issues, that I know of.

Running on computational cluster - CentOS Linux release 7.9.2009 (Core)

Clade species are missing for many *.tre.bp files, e.g.,

pxbp -t example.tre -o example.tre.bp produces the following example output:

621 trees 4 unique clades found CLADE: FREQ: 0.539452 ICA: 0.08199 COUNT: 335 TREEFREQ: 0.539452 CLADE: Costus_dirzoi_98079 Costus_chartaceus_19269 Costus_zingiberoides_98162 FREQ: 1. ICA: 1. COUNT: 621 TREEFREQ: 1. CLADE: Costus_zingiberoides_98162 FREQ: 0.243156 ICA: -0.08199 COUNT: 151 TREEFREQ: 0.243156 CLADE: Costus_chartaceus_19269 FREQ: 0.217391 ICA: -0.08199 COUNT: 135 TREEFREQ: 0.217391 TSCA: 1.24597

When pxbp is run on the exact same files on my local computer (MacOS V. 12.2.1) where phyx was installed with homebrew, it works as expected.

josephwb commented 1 year ago

@kateuckele Can you provide the input files?

kateuckele commented 1 year ago

Sorry for the delay! Here is an example of a .tre file and its corresponding output .bp file test3.rooted_trees.tre.bp.txt test3.rooted_trees.tre.txt

josephwb commented 1 year ago

Thanks. I'm not sure exactly what the problem is. Is it that the clade members are not listed in the first clade? What are your expected results?

josephwb commented 1 year ago

Hrm. Okay, running it as you did I got the same results as you did, but when I re-ran it with the verbose flag (-v) it seems to generate what you want:

pxbp -t test3.rooted_trees.tre -o example.tre.bp -v
josephwb commented 1 year ago

If this solves things, then perhaps you have 2 different versions of phyx, with each one having a different default value for verbosity.

Please try re-running the example files using the verbose flag and let me know if things are working correctly.

kateuckele commented 1 year ago

I don't think that worked, but you are right that I'm using two versions of phyx. I'm using version 1.2 on my local computer, and I'm using 1.3 on the remote machine. I've attached a few output files for you. I've identified a potentially new issue that the v.1.3 output doesn't match the v.1.2 output (w/ same input tree files). While the verbose v.1.3 version matches the non-verbose v.1.3 version, both of the v.1.3 versions do not match the v.1.2 version. As one example, the v.1.3 versions produce 6 unique clades, and the v.1.2 version produces 4. Also, in the v.1.3 verbose version, the taxa included in the first clade is still missing (which is the initial issue that I wrote about).

"test3.rooted_trees.tre.bp.v.1.3.txt" is the output from running pxbp -t test3.rooted_trees.tre -o test3.rooted_trees.tre.bp on the remote machine with pxbp 1.3 test3.rooted_trees.tre.bp.v.1.3.txt

"test3.rooted_trees.tre.bp.verbose.v.1.3.txt" is the output from running pxbp -t test3.rooted_trees.tre -o test3.rooted_trees.tre.bp -v on the remote machine with pxbp 1.3 test3.rooted_trees.tre.bp.verbose.v.1.3.txt

"test3.rooted_trees.tre.bp.v.1.2.txt" is the output from running pxbp -t test3.rooted_trees.tre -o test3.rooted_trees.tre.bp on my local machine with pxbp 1.2 test3.rooted_trees.tre.bp.v.1.2.txt

Here is the input tree file test3.rooted_trees.tre.txt

josephwb commented 1 year ago

Yikes. Okay. It sounds like v1.3 is broken, almost certainly by me...

I will try to get to this soon. Thanks for pointing it out.