Cell Ranger Arc on 10x Multiome data ERROR: there was no I2 read FASTQ and we were unable to read a 16-base barcode from the FASTQ header

Question

I am trying to run cellranger-arc count on my FASTQ files I got back from my 10x multiome sequencing run on a NovaSeqX (I submitted it to a core and they demultiplexed it - I already reached out but reply will be slow due to holidays)

I'm getting this error when I run cellranger-arc count AND when I upload to 10x cloud analysis:

[error] Pipestance failed. Error log at: 20231122GW18EC/SC_ATAC_GEX_COUNTER_CS/SC_ATAC_GEX_COUNTER/_ATAC_MATRIX_COMPUTER/MAKE_ATAC_SHARDS/fork0/chnk0-u909384e43f/_errors

Log message: Unable to read barcode sequence for read ID LH00416:29:22GF7TLT3:4:1101:1259:1032 1:N:0:NGCGTTTC: there was no I2 read FASTQ and we were unable to read a 16-base barcode from the FASTQ header. Make sure that the flow cell was demultiplexed correctly.

I do not believe my error is with my coding because I simply uploaded it to the 10x cloud analysis software and got the same error. I messed around with naming and loading the files in different ways but nothing worked. I don't have the original output, I only have the demultiplexed files.

I just want to troubleshoot while I wait (a very long time) to get a response from the core.

Answer 1

I had same issue. In my case the problem was that since Index 2 (I5) was not used in demultiplexing (specifically Illumina's bclconvert), it did not write the bases into a fastq file (*_I2_001.fastq.gz).

I solved this by changing:

OverrideCycles,Y50;I8;N24;Y49 -> OverrideCycles,Y50;I8;U24;Y49

where 'U' stands for UMI read. Also I had to add:

TrimUMI,0

So that the bases from these cycles were written to a fastq file (*_I2_001.fastq.gz) that the Cellranger-Arc needs.

Since you received already demultiplexed fastq-files, do those contains files from the second index *_I2_001.fastq.gz? If not, then you could ask the core department to rerun demultiplexing with these changes, assuming they are also using bclconvert.

If they are using bcl2fastq then I believe they should use:

Y50,I8,Y24,Y49 as the base mask, if I remember correctly.