I'm building an open-source Python library (samplesheet-parser) that validates and merges Illumina SampleSheet V1/V2 files before BCLConvert runs. When implementing cross-sheet index collision detection, I hit an ambiguity I can't resolve from the spec alone: When Lane is omitted entirely from [BCLConvert_Data], does BCLConvert 4.x:
(a) Treat all physical lanes as one pool and apply [BCLConvert_Settings] once, or (b) Fan out per-lane internally and apply settings independently per lane?
Doc #1000000004084 says "omitting Lane means all lanes" but doesn't clarify the Settings interaction. My merger currently treats omitted Lane as a single virtual pool for collision detection — flagging two samples from different project sheets sharing an index as a collision. If BCLConvert fans out internally, that's correct. If it truly treats them as one pool, I may be over-flagging. Two secondary questions:
Does AdapterBehavior (v4+) override or combine with AdapterRead1/AdapterRead2 when both are present? Are per-lane [BCLConvert_Settings_N] sections officially supported in v4.x?
Any insight from people running BCLConvert in production on NovaSeq X would be really valuable.
0 answers
No answers yet.
Log in to answer this question.
If you have access to a local install of
bclconvert(which you must), you may need to test some of this yourself (or email Illumina tech support and ask).Does your facility run a Nova X with a single giant pool across all lanes? That is likely not how most facilities run the customer samples (which means that lane specifications are always needed).
Since you are writing the parser, you can easily create multiple lane/project specific samplesheets and parallelize the processing, if you have a cluster/hardware available.
Any new software project is welcome but what is the general need for this parser. Sounds like this is more of an internal need. Are your customers submitting samplesheets in
bcl2fastqformat that you need to deal with. Normally all the relevant sample info should be coming in via sample submissions/LIMS, which would then be used for generation of samplesheet(s).Far as I know,
bclconvertsupports all sequencers so a single samplesheet should work for everything.Thanks, GenoMax! The validator, converter, and CLI are broadly useful for anyone working with BCLConvert, catching format errors before a run, migrating V1 to V2 for NovaSeq X, or adding samplesheet validation to Nextflow/Snakemake pipelines.
The merger specifically targets sequencing cores by merging per-project sheets from multiple PIs to catch cross-project index collisions before they cause demultiplexing failures.
On the Lane question, I'll test empirically and report back here. Will also reach out to Illumina tech support.