Demultiplex Novaseq run with mixed samples containing either UDI alone or UMI-UDI

Question

[diwasri]

I have the raw basecall files from a Novaseq run that was run using the XP workflow. In each of the lanes I have 8 samples with UDI only and 8 samples with xGen UMI-UDI. I need to demultiplex the samples and generate fastq files. The command I am trying to use is

/root/bin/bcl2fastq --input-dir /pathtoNovaseqRun/Data/Intensities/BaseCalls \
               --runfolder-dir /pathtoNovaseqRun/ \
               --output-dir /pathto/output/files \
               --sample-sheet /pathtoNovaseqRun/sample_sheet.csv \
               --stats-dir /pathtoNovaseqRun/Stats \
               --reports-dir /pathtoNovaseqRun/Reports \
               --use-bases-mask Y51,I8Y9,I8,Y51 \
               --mask-short-adapter-reads 0 \
               --create-fastq-for-index-reads \
               --ignore-missing-bcl \
               --no-lane-splitting

This gives me the following error

ERROR: bcl2fastq::common::Exception: 2020-Sep-23 21:29:20: Success (0): /root/bin/bcl2fastq/src/cxx/lib/layout/Layout.cpp(419): Throw in function void bcl2fastq::layout::validateIndexLengths(const ReadMetadataContainer&, const std::vector<long unsigned int>&, bcl2fastq::layout::LaneInfo&)
Dynamic exception type: boost::exception_detail::clone_impl<bcl2fastq::common::InputDataError>
std::exception::what: Barcode lengths in the sample sheet do not match those in --use-bases-mask

A small excerpt from the samplesheet is linked

How can I demultiplex this set of data?

Thank you!!

Answer 1

[GenoMax]

You have to use a specific method for IDT xGen UDI-UMI data demultiplexing. It is described in this tech note. You can create a samplesheet for the UDI only samples and demultiplex them separately.

[diwasri]

Thank you. So, to clarify, I will need to have separate sample sheets for the samples with UDI only and for those with the xGen UDI-UMI and then process them separately.

For the UDI only samples, can I use the code originally mentioned with the exception of the --use-bases-mask changed to Y51,I8,I8,Y51

[GenoMax]

You should use Y51,I8n*,I8,Y51. Assuming index 1 has been sequenced to 17 bp,