FerriolCalvet
commented
1 year ago Hi Katharina,
First of all thank you for trying geneidx and for reporting these issues. I will try to provide solutions to them and let me know whether it works.
-
Regarding pulling the singularity images, I have not found this issue before, but after checking the current singularity versions I see that it has been updated many times since version 1.0.0, so you could try to update it and then it should work fine. https://docs.sylabs.io/guides/3.0/user-guide/installation.html In this link they detail how to install the latest releases since using
apt install singularityis installing only the first version. -
For this issue with geneidx/Nextflow trying to write in a directory where it should not write. Based on our implementation all the output is written inside the output folder so there should not be any process trying to write in the input directory. It could be that when the pipeline tries to mount the directories for getting the input there is some writing involved, but I don't really know how Nextflow manages this. In one of the previous issues, there was a similar error regarding permissions and containers and it was solved by adding the
--bindoption to the executed command, you could try to do it this way and let me know if it works or not. I will look for other possible solutions. -
Regarding this last error with the taxid, I have run the pipeline with the test dataset and this taxid and it worked without any problems selecting a set of proteins and also using the parameters from Drosophila melanogaster that are the closest ones. By looking at the output you uploaded, it looks like there is still an error with the mounting of the docker images. I would say that this is more likely to be the reason for the error. Again, if you could try with the
--bindoption it could contribute to solve it. In case you prefer to provide the set of proteins to use, you can add the--prot-fileoption to indicate the proteins to use. It should be compressed (.gz) and for getting a proper naming of all the files it would be ideal to have it as (.fa.gz). Also for the genome fasta file.
In order to test these problems with containers I would recommend to start by running the pipeline only indicating the taxid, since there are already a default genome and output directories.
nextflow run main.nf -profile docker --taxid 7240If it does not work then I would be more convinced that the problem is due to the connections with the containers, if it works try adding the genome and your preferred output directory and try again.
Thank you again for trying geneidx and reporting these issues. Also let me know if it works or you find other problems.
Ferriol
KatharinaHoff
Hi,
I tried to run geneidx. With singularity, I failed to pull the image - but possibly that's an issue on my own server.
Here's my output of singularity --version (installed via apt-get on Ubuntu ):
With docker, I execute with root permissions. The image can be pulled. The problem is that apparently the container tries to write in places where the input data sits.
Here is my command:
Here is my error message:
It is rather clear how to fix (don't pipe the unpacked genome to any other place but the output directory).
Next, I copied the gzipped genome into a folder where root has writing permissions and tried again. I am using reference species Drosophila simulans with taxon id 7240. That is indeed the taxon id in NCBI Taxonomy
Here is my new call:
It fails to find the taxon for some obscure reason. There are most definitely D. simulans proteins at NCBI for this taxon, I have the protein set on my harddrive, too. I just don't know how to start the pipeline with a local protein set. Or maybe it finds the proteins, but s.th. goes wrong looking for geneid parameters for this taxon?
Error messages:
Best wishes,
Katharina