Sequencing data · What you can do here
Run an nf-core pipeline without a cluster
Run standard nf-core pipelines locally using Nextflow. Validate your samplesheet against the pinned pipeline schema, execute on a dedicated single-machine job, and inspect MultiQC reports without managing cluster queues.
What this involves
nf-core provides over one hundred community pipelines built on Nextflow, packaging JSON-schema-validated parameters and modular toolchains for common sequencing assays. Doing this work properly does not involve writing custom workflow code. It requires selecting the appropriate pipeline for your scientific objective, constructing a strictly compliant samplesheet, pinning exact pipeline release revisions, and establishing a valid configuration.
On this platform, pipelines execute as isolated jobs using Nextflow's local executor on a dedicated machine. The system supports jobs sized up to 32 cores, 256 GiB of RAM, and 24 hours of runtime. Process dependencies are managed through conda environments built by mamba on first use and cached locally. This standalone setup comfortably accommodates test profiles, targeted gene panels, amplicon sequencing, single genome assemblies, or standard RNA-seq cohorts, without requiring cluster scheduling or software installation on your part.
Because the execution machine operates as a single host rather than a distributed compute cluster, extremely resource-intensive projects—such as 30x whole-genome sequencing cohorts or 200-sample RNA-seq experiments—exceed local limits. For these, the agent validates your inputs, parameters, and samplesheet, and generates the exact pinned Nextflow launch command to run on your own cluster or institutional infrastructure.
How it goes
- 1Select pipeline and inspect schemaSearch available nf-core workflows by your scientific objective and inspect the parameter schema for a pinned release. Verify the required parameters, accepted inputs, and release date before configuring runs.
- 2Construct and validate the samplesheetCompile sample metadata and FastQ file paths into a CSV samplesheet, validating every entry directly against the pinned pipeline's input schema to check required fields, formatting patterns, and experimental design constraints.
- 3Run pipeline test profileLaunch Nextflow with the test profile and local conda execution on the dedicated machine to build cached mamba environments and confirm networking and pipeline parameters before queuing experimental data.
- 4Execute full pipeline runExecute the pipeline with pinned revisions, explicit parameters, and absolute input paths pointing to files located in the upload directory, using Nextflow's local executor with job resumes enabled.
- 5Inspect MultiQC and pipeline outputsExamine the MultiQC quality control report located in the results directory, checking alignment rates, inferred strandedness, ribosomal RNA fractions, and execution metrics recorded in pipeline_info.
Start it here
Ask in plain language. The agent works on a Linux machine of its own — it installs and runs what the job needs, keeps every file, and reads the results back with the numbers explained. Nothing to set up, and a long run continues after you close the page.
What to watch for
- Never omit the -r flag; running without a pinned revision defaults to the newest branch, breaking repeatability and methods citations.
- Do not pass a generic 'unstranded' flag simply because strandedness is unknown; set 'auto' so Salmon can infer it, as misdeclaring unstranded on stranded preps quietly distorts counts.
- In nf-core/rnaseq, the sample column serves as the merge key; giving the same sample name to distinct biological subjects will merge their reads silently.
- Avoid using outdated AWS iGenomes gene annotations; supply explicit --fasta and --gtf paths from defined GENCODE or Ensembl releases instead.
- Remember that parameters defined inside a custom infrastructure configuration file (-c) will not override default pipeline parameters in nextflow.config.
- Do not assume nf-core/rnaseq performs differential expression; it outputs quantification matrices and quality control, requiring nf-core/differentialabundance for statistical testing.
What it runs
Questions
Do I need to install anything to run an nf-core pipeline without a cluster?
No. The research agent does this work on a Linux machine of its own — it runs Nextflow there, keeps the files, and reads the results back to you. Long runs continue if you close the page.
Can I run container engines like Docker or Singularity on the platform machine?
No. The local execution machine does not support container engines, so it uses the conda profile with mamba-built environments. If you require Singularity or Docker profiles, the platform generates the exact pinned command line and parameter file for your external infrastructure.
What size runs fit within the local job limits?
A job provides at most 32 CPU cores, 256 GiB of memory, and 24 hours of execution time. This fits built-in test profiles, typical RNA-seq runs, amplicon sequencing, targeted gene panels, and single genome assemblies, but not large cohorts or deep whole-genome sequencing.
How are single-end reads configured in an nf-core/rnaseq samplesheet?
You must include the fastq_2 column in the CSV file and leave the corresponding cells empty. Completely removing the column causes the run to fail schema validation.
What happens if a task fails due to memory exhaustion?
nf-core pipelines automatically escalate allocated memory on subsequent retries. A task that fails on its initial attempt and finishes successfully upon retry is normal behaviour.
Written from the method the research agent follows for this work and the documentation of the tools listed above. Statements citing a figure that does not appear there are removed before publication. Corollary is a non-commercial research demonstration; results are not for clinical use.