PF2 pre-release 11 is looking for the wrong tree file following estimation of RAxML tree

[marekborowiec]

After setting up an analysis with the --ml-tree option I get this error:

ERROR    | 2016-01-30 02:44:43,993 | A program that Partitionfinder uses failed. Output follows, in case it's helpful for finding the problem
ERROR    | 2016-01-30 02:44:43,993 | 
ERROR    | 2016-01-30 02:44:43,994 | . command-line: /datadrive/partitionfinder-2.0.0-pre11/programs/phyml --run_id GTR+I+G -b 0 -i ./analysis/phylofiles/58ab28da375c2cc3efefce8d102a994b.phy -u ./analysis/start_tree/filtered_source.phy_phyml_tree -o lr -m 012345 -f e -v e -a e -c 4 --min_diff_lk_local 0.01 --min_diff_lk_global 0.01 --no_memory_check . The file './analysis/start_tree/filtered_source.phy_phyml_tree' doesn't exist. . Type any key to exit.

It looks like the tree PF2 should be looking for is ./analysis/start_tree/RAxML_result.BLTREE. When I move this RAxML tree to the input folder and re-run the program with it as user-supplied topology it works fine.

[roblanf]

Hi Marek,

Sorry for the slow reply - I've been on paternity leave.

Can you tell me the commandline you are using to run this analysis. My guess from the output is that you did not specify the --raxml flag, is that right? I need to add in a condition for what the program should do when you want to use the --ml-tree option in this case.

Rob

[marekborowiec]

Correct, I didn't use --raxml. I figured out later that using --raxml in combination with --ml-tree works. This is a minor bug, but, like you say, a condition would be nice to make sure that the program doesn't trip following estimation of ML tree.

Best, Marek

[marekborowiec]

I'm now working with pre-release 13 and still have a similar problem, although now there is no error. I want to look for MrBayes models using the greedy algorithm and so cannot use the --raxml which limits models to variations of GTR. I run PF with:

python ~/Phylo-software/partitionfinder-2.0.0-pre13/PartitionFinder.py --ml-tree -p 8 .

This estimates the ML tree fine, then hangs when starting phyml analyses. In debug mode, I can see that the RAxML tree is correctly estimated ant saved into ./analysis/start_tree/RAxML_result.BLTREE. However, a few lines below PF declares:

starting tree with branch lengths is here: ./analysis/start_tree/filtered_source.phy_phyml_tree.txt

In the end, PF starts eight phyml runs trying to use that tree and hangs. After I rename (or copy) the RAxML result tree to filtered_source.phy_phyml_tree.txt and restart the analysis, everything runs smoothly.

[roblanf]

Hi Marek,

Thanks for this, I'll take a look as soon as I can.

Rob

On 23 April 2016 at 08:34, Marek notifications@github.com wrote:

I'm now working with pre-release 13 and still have a similar problem, although now there is no error. I want to look for MrBayes models using the greedy algorithm and so cannot use the --raxml which limits models to variations of GTR. I run PF with:

python ~/Phylo-software/partitionfinder-2.0.0-pre13/PartitionFinder.py --ml-tree -p 8 .

This estimates the ML tree fine, then hangs when starting phyml analyses. In debug mode, I can see that the RAxML tree is correctly estimated ant saved into ./analysis/start_tree/RAxML_result.BLTREE. However, a few lines below PF declares:

starting tree with branch lengths is here: ./analysis/start_tree/filtered_source.phy_phyml_tree.txt

In the end, PF starts eight phyml runs trying to use that tree and hangs. After I rename (or copy) the RAxML result tree to filtered_source.phy_phyml_tree.txt and restart the analysis, everything runs smoothly.

— You are receiving this because you commented. Reply to this email directly or view it on GitHub https://github.com/brettc/partitionfinder/issues/85#issuecomment-213612996

Rob Lanfear School of Biological Sciences, Macquarie University, Sydney

phone: +61 (0)2 9850 8204

www.robertlanfear.com

[roblanf]

Done (sorry it took a long time!): https://github.com/brettc/partitionfinder/commit/f6dd6eca5da4fdb080ae17ddb7c2a46feecebb74