Share a MultiQC report
MultiQC does the aggregation nobody wants to do by hand: forty FastQC outputs, alignment stats, duplication metrics, all collapsed into one interactive page where an outlier sample is visible at a glance.
Then the page is a 30 MB HTML file on a cluster filesystem, and the collaborator who needs it is at another institution with no account on that cluster. So it gets emailed, bounced for size, uploaded to a file service, and discussed in a mail thread where nobody can tell which sample anyone means.
Publish it
multiqc . -o qc/
curl -fsS -X PATCH \
"https://commareports.com/api/v1/reports/$COMMA_REPORT_ID" \
-H "Authorization: Bearer $COMMA_API_TOKEN" \
-H "Content-Type: application/json" \
-d "$(jq -n --rawfile html qc/multiqc_report.html \
--arg title "QC — $RUN_ID — $(date +%F)" \
'{title: $title, html: $html}')"
Or drag multiqc_report.html into the app.
The plots stay interactive at the link — scripts run inside a sandboxed
iframe (allow-scripts, no allow-same-origin).
One report id per run batch. PATCH after a re-run keeps the
collaborator's link working, with the previous QC still available as a
revision — which is the question that comes up after re-basecalling.
What the URL changes
- Annotations on the sample. "S14 is the low-input library, expected" pinned to the point on the plot. See commenting on HTML.
- Collaborators need no cluster access. A link, in a browser.
- Run-over-run comparison. See revisions and diffs.
- Private by default. Sample identifiers are often more identifying than they look — see the sharing model, and use a password-protected link for external collaborators.
Limits
- HTML body: 5 MB. A MultiQC report over hundreds of samples exceeds
this comfortably. Use
--flatfor static plots,--interactiveonly where it earns its size, and split by module or by batch — or publish the summary and attach the full report as an asset. - 60 requests/minute per token.
Try it
Comma is free — unlimited reports, unlimited commenters, unlimited revision history.
Related
- Nextflow run reports — the pipeline around it
- Jupyter notebooks · Quarto
- ydata-profiling · Publish from CI