As of January 2024 we have changed the tool we use for creating FASTQ files for Illumina sequencing from the now deprecated bcl2fastq to their current recommended tool, BCL Convert.
For most library and index types there will be no noticeable difference in the output from BCL Convert from what you are used to. There are however some cases where there is a significant difference; these are where the kit produces UMI reads.
The library types affected are:
The index types affected are:
Instead of a third read, where in most cases the second read of the three is the UMI read, there will be only two reads containing the regular reads. The UMI bases will now appear in the read header instead, much as the index bases have always done.
A full comparison of the changed between the two tools is available on the Illumina web site.
Below is a read from a lane of 10x Multiome ATAC sequencing created with bcl2fastq:
Its UMI read is this:
Processing the same lane with BCL Convert gives the following for the same read:
As you can see, the BCL Convert read is identical to the bcl2fastq read one except that the UMI is present in the header. There is no UMI read and thus no quality scores for the UMI bases, but this is consistent with the index bases in most cases where we do not provide index reads.
We cannot know about every tool that uses the UMI reads and how any given one might handle the new format. We've tested 10x's cellranger and that seems to work with files produced from BCL Convert without issue.
On the subject of 10x tools, we have tested them without providing FASTQ files for the index read and the output is identical, with or without the read or reads. As such, we will no longer be producing FASTQ files for the index reads for 10x library types.