How to run barcodeSwapping and ChimericReads with no h5 info?

[mortunco]

Hi,

First of all thank you for coming up with such a critical procedure that would effect almost all single cell study.

Background: We implemented 10x to our in-house method that gives read out as RNAs to make "single cell" version of our method. Because our method contains couple additional multiplexing steps and mainly based on non-coding regions we simply couldn't implement cell-ranger standardized method pipeline.

Problem: Please see the sample dataframe below. We observe same moleculer UMIs from the barcode(droplet) are pointing different transcript fragments which doesnt make sense. Instead of some arbitrary filtration, i think this problem is similar to problems in "Removing Swapping Events" section. But I see that most of your tools are built to process standard cell-ranger output (barcode, gene, matrix.mtx or h5) which makes becomes a little problematic for us.

Attempt: We successfully generated a dataframe like below. rindex and cindex are similar to transcript~ gene info. We were able to create barcode, gene tsvs and matrix.mtx files. Using these I created sce object with read10xCounts. Droputils allows me to run empty drops but for Swapping i need mol.info, but i dont know how to convert my data into h5 format.

cellb	molb tindex	n
AAAAAAGGTATGGTTC	TAAATCAAAAAA	A	5
AAAAAAGGTATGGTTC	TATCTCAAAGAA	B	10
AAAAAAGGTATGGTTC	TCACTCAAAGAA	C	1

Question: Pretty much, I would be glad to hear any ideas about problem or h5 conversion.

Thank you very much for your time,

Tunc.