a question to the sig group, is Seqera platform ne...
# meta-omics
o
a question to the sig group, is Seqera platform necessary requirement to run chained pipelines for the meta-omics project? @Evangelos Karatzas
e
Good question, I haven't followed the meta-pipelines at all. Lemme make a fast check
šŸ¤– Pipeline composition (the "meta-pipeline" feature) is core Nextflow, so it works from the plain CLI for everyone. It is not a Seqera Platform feature. Platform only has a separate approach, chaining, built from its automation tools.
Limits • It's marked Experimental (docs/workflow-typed.mdx L67–70) and may change. • The calling script and the included pipeline must both turn on static typing (nextflow.enable.types = true). • Only local paths work. Including a remote pipeline is deferred (ADR adr/20260608-pipeline-composition.md L21, L161). You bring the pipeline in yourself with a git clone or a submodule. • nf-core pipelines aren't ready for it yet: they're untyped and have no output block. The Barcelona hackathon proof of concept used trimmed copies.
m
We haven't rolled out static typing yet across all nf-core workflows which is a necessity for chaining pipelines with the new pipeline composition method, but there's a blog post coming soon about the nf-core modules will change. Otherwise if you do have platform, there's the node solution Phil presented, or if you're working on HPC there's my nf-cascade method ( nextflow-in-nextflow; see here for an untested meta-pipeline ). These last two don't need the latest syntax
p
This is the Node-RED thing: seqeralabs.github.io/node-red-seqera
See #C07ACNLELR2 for more discussion
party this 1
o
yes, exactly, I am trying to figure out if we can use nf-meta for it, but
nf-meta run
can only drive a chain by shelling out to
nextflow
directly against whatever's on the machine running it — e.g. our slurm. It has no concept of a Seqera Platform API call. Right, @Julian Flesch?