Hi all,
Many Nextflow pipelines are not fully compatible with the Seqera platform. By “fully compatible” I mean that when adding one of these pipelines to Seqera and clicking “launch” there is no page to specify inputs, outputs, or other parameters. Some of these pipelines are designed by institutions that actively use dedicated HPC systems.
I would like to understand how to make these pipelines compatible with the platform. What steps and considerations should be kept in mind for this process?
My first recommendation would probably be to see if there is a different pipeline that is already Seqera compatible. You can start by looking at Seqera Pipelines which is a curated list of pipelines (both nf-core and not) that meet standards for being portable and deployable to any infrastructure. From searching in that list, you can see that nf-core/chipseq also exists might meet your needs for this type of data anlysis.
If you can’t find an alternative pipeline that in nf-core or Seqera pipelines, then you’ll have to go through the process of manually verifying if the pipeline was written to be portable by testing it on your infrastructure. Sometimes pipelines are written with a lot of hard-coded variables and things specific to their infrasturcture and you won’t find those until you test it out.
Once you confirm that the pipeline runs, you can very easily generate the nextflow-schema.json file that Seqera Platform uses to generate the beautiful launchpad interface. Easiest way to do that is with the nf-core cli and the command nf-core schema build. See these docs for details.