Closed LehmannN closed 4 years ago
Dear @LehmannN ,
This is really surprising, I automatically check that Daijin runs correctly whenever there is a commit. May I ask which version of TransDecoder you are using? On my workstation Universal
is explicitly listed as a valid option (with conda
, transdecoder=5.5.0=pl526_2
).
I found that I am running TransDecoder-3.0.1 so that might explain why. I have it installed locally and through the conda environment but the default program that runs is the one installed locally... I am not sure how to fix that but will have a look.
I realized that in fact I had no installation of transdecoder
in conda, only a local installation. I have been fooled by running mikado inside the conda environment, but since the tool was not in the environment, daijin found it in /usr/bin/. Is there a reason why you don't have transdecoder
and mikado
in the environment.yml
file ?
Dear @LehmannN ,
Within Daijin, I try to keep the different tool requirements separate, also because there are tools that are incompatible with each other (e.g. I think that tophat2
requires python2, which would make it incompatible with Mikado itself).
For this reason, Daijin fully implements the --use-conda
switch present also in Snakemake. Briefly, if the switch is present on the command line, Daijin will create a separate Conda environment for each required tool, including TransDecoder.
I hope this helps.
Kind regards
Ok, thanks for the explanations !
Hello !
I have another comment regarding the daijin pipeline. When coming through the
TransDecoder.LongOrfs
step in thedaijin mikado
pipeline, it suddenly stops with the error "Sorry, Universal is not currently supported or recognized.". So I just ran transdecoder out of the daijin pipeline without the-G Universal
parameter, but I wonder it there is a way to modify the command line ran bydaijin
? I am using the Mikado conda environment btw and we get the same error forTransDecoder.Predict
.Thanks again help and comments.