Generate consensus reads without using UMIs

[SPPearce]

Is there a way to generate consensus reads, without using the UMI information at all?
I'd like to see how much difference using the UMIs is actually making.

[nh13]

I haven't tried it, but if you stuck in some dummy values into the RX tag before grouping (RX:Z:A for single UMI and RX:Z:A-A for duplex), you could then run through grouping and consensus calling and the UMIs would effectively provide no additional information. For non-duplex this should work. For duplex-sequencing, the UMIs help identify the original source strand of the molecule, so dummy values would would cause all reads to be on the top-strand.

You could also run picard MarkDuplicates with the TAG_DUPLICATE_SET_MEMBERS option, then copy the ID tag to the MI tag prior to consensus calling.

Let us know what you find!

[SPPearce]

• For duplex-sequencing, the UMIs help identify the original source strand of the molecule, so dummy values would would cause all reads to be on the top-strand.

Hmm, I had the feeling the reads were always assigned to the forward stand in duplex mode?

[nh13]

See /A and /B https://github.com/fulcrumgenomics/fgbio/wiki/Developer-Note:-Tracking-Reads-through-Grouping-and-Duplex-Consensus-Calling

Forward strand of the genome? Or forward strand of the original source molecule? It's an overloaded term for duplex sequencing...

[SPPearce]

I think that the best way to actually do this is to up the allowable edit distance to something larger than the umi_length, which will make all the UMIs to be grouped into one.