In rare situations where lanes (or lane subsets) contain projects with differing index sequence templates within a sample sheet, currently the Fastq generation pipeline is unable to handle this directly and stops with an error (pre-empting failures from bcl2fastq, which it is assumed also cannot handle this situation).
The proposal is to extend the pipeline subsets to allow projects to be specified as well as or instead of lanes, which would circumvent the issue of differing index sequence lengths (this is effectively equivalent to the official Illumina solution of using separate sample sheets for each project, and running the bcl2fastq step once for each).
The most significant impact in these cases would be that the "undetermined" Fastqs could no longer be useful combined from the runs (as reads from one project would be undetermined for another).
In rare situations where lanes (or lane subsets) contain projects with differing index sequence templates within a sample sheet, currently the Fastq generation pipeline is unable to handle this directly and stops with an error (pre-empting failures from
bcl2fastq, which it is assumed also cannot handle this situation).The proposal is to extend the pipeline subsets to allow projects to be specified as well as or instead of lanes, which would circumvent the issue of differing index sequence lengths (this is effectively equivalent to the official Illumina solution of using separate sample sheets for each project, and running the
bcl2fastqstep once for each).The most significant impact in these cases would be that the "undetermined" Fastqs could no longer be useful combined from the runs (as reads from one project would be undetermined for another).