Support/distinction of other linked-read technologies

[pdimens]

Hello, I'm interested in incorporating this tool into Harpy, which is specific to haplotagging-style linked reads. The documentation provides the option for 10x and stFLR protocol types, but not other varieties like haplotagging or TELLseq.

My question is: is there an appreciable difference of what the two protocols alter within physlr and would it be possible to add support for haplotagging-style barcodes? The haplotagging barcodes are still provided in the BX:Z tag, but are in the format AxxCxxBxxDxx where each xx is an integer between 0-99 e.g. A02C96B34D51.

Sort of related to that, is that deconvolved barcodes may feature a hyphenated ending (e.g. BX:Z:<barcode>-1) or an additional MI:i: tag of a simple integer indicating the unique molecule those barcodes are said to be associated with (e.g. MI:i:41 and MI:i94122). Is there value in prioritizing the MI:i tag, if present, over the BX:Z tag?

[lcoombe]

Hi @pdimens,

Sorry for the delay in getting back to you!

As long as the 'barcode' is in the format. BX:Z:<barcode>, it will be recognized by the parsing software in physlr.
Any distinct string for the "barcode" section will be recognized as a different barcode - but any MI:I tag isn't recognized. So if you want to consider different barcode+molecule tags as distinct 'barcodes', then you'd have to alter the barcode part to contain that identifier - that is what we have done in the past when, for example, we had different chromium libraries that we wanted to use together.

I hope that helps - thank you for your interest in Physlr!
Lauren

[pdimens]

Does physlr distinguish when a barcode is invalidated? For example, when any of the segments of a haplotag barcode feature 00, e.g. A51C00B18D91

[lcoombe]

No, it doesn't do any checking of the barcode values - it will just take the full string (from the expected barcode format) as the barcode. So, if you want some barcodes to be ignored (if they are invalid, etc), you'd need to do some pre-filtering.