Fastqs different lanes

Question

[samuel]

Can anyone please confirm the following.

• You should generate sam/bams with read groups for downstream analysis (GATK)

• If you have fastqs from the same sample run over different lanes of the flowcell these would have different read groups

• Therefore you should keep the fastqs seperate, align whilst adding read groups and then merge after alignment.

• By concatenating fastqs before the alignment step you would lose the read group information.

Answer 1

[swbarnes2]

Fastqs don't have read groups. The same sample split across different lanes can all be catted together as fastqs.

[samuel]

I'm aware fastqs don't have read groups and that we can add them at the alignment stage. I was wondering as to whether we need to preserve lane information for downstream applications such as MarkDuplicates i.e. when marking optical duplicates?

Answer 2

[benformatics]

Follow the GATK standards probably:

https://gatk.broadinstitute.org/hc/en-us/articles/360035889471-How-should-I-pre-process-data-from-multiplexed-sequencing-and-multi-library-designs-

Note that for the purposes of marking duplicates you would merge the lanes as the libraries themselves contain PCR duplicates across lanes.