New release btw -> 1.7.1 -- including MQC SAV, bugfixes, workflow outputs 😉 and strict syntax + schema 2.7.0 🙂

Congrats

Hello, thanks for the work on demultiplex! The group is starting to process raw data from NovaSeq6000 with dual adapters and UMIs, so I am in the painful process of learning conversion cbcl 🥲 My latest issue was the fastq files are not transferred from the work to the output directory, but the pipeline is successfully run (fastqc files, etc everything good). This was on demultiplex 1.7.1 with nextflow 26.04.0. I solved it by downgrading back with

NXF_VER=25.10.5 nextflow run

[main] ERROR nextflow.script.WorkflowMetadata - Failed to invoke `workflow.onComplete` event handler
groovy.lang.MissingMethodException: No signature of method: sun.nio.fs.UnixPath.rightShift() is applicable for argument types: (String) values: [XXXXXXXXXXXXX_S1_L001_R1_001.fastq.gz]
Possible solutions: leftShift([B), leftShift(java.io.InputStream), leftShift(java.lang.Object)

Hello, I hope you are doing well. I am trying to use Nextflow / nf-core to demultiplex a 10x cell-hashing (HTO) library. We have 10 hashing barcodes (15 bp) corresponding to 10 sample groups (mouse × condition × timepoint), and the hashing FASTQs are split across two lanes (L001/L002; paired-end). We validated the barcode is in R2 (starts at position 11), so we used R2 read structure 10S15B65T and ignored R1 (28S). We ran nf-core/demultiplex with --demultiplexer fqtk (v1.7.1, docker). Running both lanes together repeatedly failed with an error like “observed_bases 30 vs expected_bases 15”. Running lane-by-lane with per-lane tar.gz inputs worked for both lanes, and we combined the outputs by concatenating the per-sample *.fastp.fastq.gz files across lanes. Could you help confirm: 1. Is lane-by-lane the recommended/expected approach for fqtk with multi-lane inputs, or is there a correct way to run both lanes together? 2. Is concatenating demultiplexed FASTQs across lanes the right way to combine outputs? 3. Does this pipeline provide (or is there a recommended way to generate) a cell-barcode → sample assignment table for downstream integration? Our end goal is to label cells for GEX/VDJ analyses and ultimately build Dowser lineage trees with nodes labeled by GEX-derived cell type. Thanks a lot for any guidance you can share!

looks interesting

How does this relate to the samshee python tool integrated already?

Hej, I have been working more and more on #C06PHJM0Y3D lately, and I do believe #CL88J906S and #C06PHJM0Y3D would be perfect for #C08J1R63EUQ and similar structures. So I will probably get more involved into it in the future

I can see a No space left on device in CI for https://github.com/nf-core/demultiplex/pull/399. Totally unrelated to this PR, so I will force merge, and fix it in a subsequent PR

Great - I'm on the latest template merge btw

Takes a while, as I cannot just fully dedicate time on it but will be there soonish

Oh, I am on it too, sorry, I should have warned beforehand

No prob, getting it done is the goal 🙂

Hi! I have been working on this PR with Maxime to add seqinspector samplesheet generation. I have added the flowcell column to the samplesheet.

We merged in https://github.com/nf-core/demultiplex/pull/402 even if not fully finished as it's not breaking anything + fixing tests from https://github.com/nf-core/demultiplex/pull/407

Hi all, We're running an internal minimal demultiplexing workflow based on the nf-core

BCLConvert

module (same module code as

nf-core/demultiplex

) on Seqera Platform + AWS Batch. We're processing a large Illumina run (~870 GB). Interestingly, the same dataset completes successfully with the official

nf-core/demultiplex

pipeline, but fails in our simplified workflow. When

scratch

and

stageInMode

are disabled, BCLConvert eventually hangs and is reset by the DRAGEN watchdog:

Hang detected - there has been no system activity for 600 seconds

No I/O activity detected for 600 seconds

The DRAGEN WatchDog service is resetting DRAGEN

When

scratch=true

and staging is enabled, the job instead fails due to local disk exhaustion:

zerolog: could not write event: write /tmp/fusion.log...: no space left on device

Fusion metrics show: download: ~869 GB available_disk: start: 1.8 TB end: 20 kB We've tested: • 500 GB and 1000 GB local disk • scratch enabled/disabled • same BCLConvert version (

4.4.6

) Since the official

nf-core/demultiplex

pipeline succeeds on the same input, I'm wondering if there are any workflow-level staging/localization steps that we're missing by using the module directly. Has anyone seen: 1. BCLConvert hanging when reading through Fusion/S3-backed storage? 2. BCLConvert requiring full localization of large run folders? 3. Differences in how

nf-core/demultiplex

handles large BCL inputs vs invoking the module directly? Any pointers would be much appreciated. Thanks!

Also, we got https://github.com/nf-core/demultiplex/pull/396 to fix https://github.com/nf-core/demultiplex/issues/395 Fix looks ok to me, but I'd appreciate another set of eyes, just in case

Update bclconvert module from 4.4.6 to 4.5.4 -> https://github.com/nf-core/demultiplex/pull/409 🙏

While we're at it, I have a question, I can see we have this module which seems to me like a waste of a process: https://github.com/nf-core/demultiplex/blob/master/modules/local/fastq_to_samplesheet/main.nf I recently replaced such a module in #C024915L4KB: https://github.com/nf-core/fetchngs/blob/25949b6fc2ade7fe26dbc5ae909b18ee170b4578/modules/local/sra_to_samplesheet/main.nf -> https://github.com/nf-core/fetchngs/blob/dev/subworkflows/local/channel_sra_create_csv/main.nf Is there really a need for such a process here, as far as I understand we are only creating a samplesheet, and do not need such a process for what can be achieved via pure groovy / channel manipulation

Do we have any blocker for a release? or can I go for it? (no worries, I won't do it today, but I can finally start working on it on monday)

I did not noticed https://github.com/nf-core/demultiplex/pull/415 I'll work on it on Monday then. Have a nice weekend all

Can we only publish the trimmed fastqs, rather than the raw ones too? This doubles the long-term storage at the moment

✅ demultiplex/test_full completed successfully! 🎉 Run Name nice_linnaeus Duration 20m 19s Status ✅ Success Tasks Completed: 79 Sep 15, 2026 at 10:50 AM

✅ demultiplex/test_full completed successfully! 🎉 Run Name stoic_perlman Duration 24m 48s Status ✅ Success Tasks Completed: 79 Sep 15, 2026 at 1:35 PM

Everyone ok for a release then? I have approvals from @Matthias De Smet @Simon Pearce and @Daniel Lundin

We have a release: https://github.com/nf-core/demultiplex/releases/tag/1.8.0 • NEW publish_raw_fastq parameter (default: false) • alert only trimmed FASTQs are published now, halving storage. Use

--publish_raw_fastq true

if you need/want raw files, raw files will still be published if no trimming. • Fixed AVITI runs without multiplexing producing empty FASTQs • Template updated to nf-core v4.1.0, bclconvert to 4.5.4, multiqc to 1.35 thanks to: @Dietmar Rieder @Anabella Trigila @Agrima Bhatt for contributions, and everyone I bothered for reviews

✅ demultiplex/test_full completed successfully! 🎉 Run Name cheeky_ride Duration 19m 12s Status ✅ Success Tasks Completed: 79 Sep 15, 2026 at 2:56 PM

depictio Depictio dashboard for demultiplex: feedback wanted Depictio (open source, from EMBL) turns the output of a workflow run into an interactive dashboard you can share, so nobody has to write a Shiny or Streamlit app for each project. Point a template at any run of a pipeline and you get the same dashboard, which you can edit (UI & YAML), version and comment. 📦 The demultiplex dashboard template (no account needed) shows run health per lane and cycle, how evenly reads are shared between libraries, and undetermined reads and possible index swaps. 🙌 We need your help! Feedback and ideas welcome in #C0AFS06G9HR (all templates and details there) or in the demultiplex GitHub discussion.

Is there someone who knows anyone on the dev team at Illumina? I have a feature request for

bclconvert

it’s proving very difficult to get through. I’d like to have them add an option for json samplesheets to replace the current csv format

Can I get a double set of eyes on https://github.com/nf-core/demultiplex/pull/435 there was a messy history, I just to make sure it's now properly merged