Will mapping Read 1 in Single Cell 3' sequencing increase mapping accuracy or not?

Question

[Zhilong Jia]

Why lots of pipeline of 3` sequencing (such as zUMI, scruff) do not map Read 1, including barcode and UMI, as PE mode mapping?

Though its low quality bases after varying-length ployT issue, will the Read 1 be helpful for mapping accurency to add this information after trimming barcode+UMI+ployT?

A related discussion about 5` sequencing is here for STAR.

"Cell Ranger has a "PE" mode that gets activated when the barcode read (read 1) is longer than 50 nt - and so it gets used for both cDNA and barcode." as indicated in the issue above.

Thank you.

Answer 1

[GenoMax]

will the Read 1 be helpful for mapping accurency to add this information after trimming

Logically, if there is enough information available in R1 for RNA part then possibly. If this would have helped then 10x would likely have recommended it as a general practice but they have not done so. Running more cycles also adds to sequencing cost. Something one will need to balance against benefit (increased accuracy, if possible).

[Zhilong Jia]

Thank you GenoMax . This is the wet lab part.

How about bioinformatic part? As the Read 1 has these sequence which can be useful. why not to map in general?

[GenoMax]

If you have longer reads for R1 you could try it out. My guess is most stick with 10x sequencing recommendations and choose shorter cycles to stay with smaller (and thus cheaper) sequencing reagent kit.

[ATpoint]

I guess this scenario would (if at all) only benefit the alignments of fragments that are anyway difficult to map due to being not very unique, and then using a short read (notably shorter than R2) with abyssmal quality to get PE data is probably (my gut feeling) not going to help but rather add uncertainty as the base calls are not reliable.