AleRax missing files in model parameter folder and DL rates

[Brenda Delamonica]

Hi,

I am trying to get D and L model parameters using AleRax. The output provides average event counts per node, but the model parameter folder is empty. The rec-model option selected is UndatedDL since transfers are not possible in my scenario. I have also tried the various --model_parametrization GLOBAL, PER-SPECIES and PER-FAMILY options and still the folder is empty for each run. Here is what I am running:

alerax --families v2_final_all_file \

 --species-tree tree6.nwk \

 --rec-model UndatedDL \

 --prefix v2_final_all_2 \

 --gene-tree-rooting UNIFORM \

 --model-parametrization PER-FAMILY

Separately, is there a way to get duplication and loss rates at nodes from the average event counts? If so, how would you recommend to do this?

Thanks,

Brenda

[Stefan Flaumberg]

Hi Brenda,

It can be a bug of the version you are using. As a workaround you can try out the bug-fixed AleRax test version (which is still under review by Noah), please use the command below to download it:

git clone -b modeltest --recursive https://github.com/StefanFlaumberg/AleRax

To get the duplication and loss rate parameters for each species tree node, please use the --model-parametrization PER-SPECIES option and then see the <output_dir>/model_parameters/model_parameters.txt file.

But I have to warn you of two things:

1. Reconstructing a species tree under the PER-SPECIES parameterization, as far as I know, has not been systematically tested (although on toy examples it works fine). You still can try it, or you can first reconstruct the species tree under the PER-FAMILY parameterization and then use the obtained tree along with the --species-tree-search SKIP and --model-parametrization PER-SPECIES to infer the node-specific parameters.

2. The model parameters cannot be straightforwardly interpreted as branch event probabilities, although such probabilities are constructed from them. So make sure you are using them correctly in the downstream analysis (please see the AleRax article supplementary materials for the math details). 

Best,

Stefan