One error about singularity pull!

[sunnycqcn]

Hello Darcy,
I am very glad to use your pipeline for predict fungi TE. I a question that I can put all genomes sequences into genomes file. Then pipeline will automatically run all genomes?
When I run the pipeline, I get the error as following. I do not know the error is from my server or the command?
I use cluster server. Because I do not have root right, I can not use docker. I installed singularity in conda3. Is it right?
Thanks,
Fuyou
WARN: Singularity cache directory has not been defined -- Remote image will be stored in the path: /isilon/saskatoon-rdc/users/fuf/BLgenome/pante/work/singularity
Error executing process > 'getRMRepeatPeps'

Caused by:
Failed to pull singularity image
command: singularity pull --name darcyabjones-default-pante-pante-v0.0.1.img library://darcyabjones/default/pante:pante-v0.0.1 > /dev/null
status : 255
message:
ERROR: pull is only supported for shub URIs

[darcyabjones]

Hi fuyou!

Sorry for the late reply.
I didn't see your issue (need to figure out email alerts).

I'm not sure about installing singularity via conda.
I suspect that the error is caused by the old version of singularity that conda distributes (<3.0) which was built before the singularity people moved hosting to sylabs.io.
Also I think singularity still needs to be installed as root even though it doesn't need to be run as root.

Can you ask your sys-admins to install singularity?
Our local supercomputing people (Pawsey) have been happily running it for a while now (until version 3.4 which introduced some features they don't like).

Alternatively you might be able to use shifter https://www.nextflow.io/docs/latest/shifter.html, if your cluster has the nersc version. I can help get a config file together for that.

[sunnycqcn]

Hello Jon,
I tried to contact our sys-admins. They can not install singularity or docker. I tried to install it in my directory. However, I can not install singularity based on the introduction. I am easy to use conda. Could you have some suggestions about conda install?
Thanks,
Fuyou

[darcyabjones]

Hi Fuyou,

It's a shame that they won't install that for you.

I'll have a look at getting a conda environment together.
But i'm quite busy for the next month or so, so it might take a while.

The issue with using conda originally was the way RepeatModeler/RepeatMasker store their databases inside of the install directory.
It made it really difficult to keep things portable.
The other issue is proprietary software (RNAmmer) which can't be bundled on the conda repositories.

RepeatModeler recently released an update to allow people to specify the database using command-line flags.
AFAIK RepeatMasker hasn't released a similar update (yet).

I'll let you know.
Honestly I'd suggest giving EDTA https://github.com/oushujun/EDTA a go.
They have a conda setup process and a much better way of dealing with nested/overlapping elements.

Sorry, It's that super busy conference time :)

Darcy

[sunnycqcn]

Hello Darcy,
I am much appreciated for your help. In fact, I have installed RNAmmer, RepeatModeler/RepeatMasker. I also can do TE analysis using each of them. However, I want to integrate all results in my fungi isolates. So I hope I can use your pipeline.
I think it is not very urgent to do. You can do it when you are free.
Thanks,
Fuyou