Single-nucleus RNA-seq processing (10x) with Pseudoaligners (Kallisto)

Question

[achamess]

I currently use CellRanger on single nucleus RNA-seq data using the pre-mRNA reference that they advise. https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/advanced/references

I want to try using Kallisto Bustools for increased processing speed. However, I haven't seen it used yet with single nucleus data, where there are a high proportion of introns, which won't be presented in the standard cDNA reference that one uses to build an index.

Is there a reference transcriptome for human that has unspliced mRNA?

Or does anyone have any references for using Kallisto with nuclear RNA?

Answer 1

[Lior Pachter]

There is a tutorial for how to do that here: https://github.com/BUStools/getting_started/blob/master/kallisto_bus_mouse_nuclei_tutorial.ipynb

This is part of a series of tutorials for how to use kallisto | bustools for a variety of single-cell RNA-seq pre-processing tasks: https://www.kallistobus.tools/tutorials

[achamess]

Thanks! Can the same cDNA/intron index be used with Kallisto for processing of bulk nuclear RNA-seq too?

[Lior Pachter]

Unfortunately it's not that simple- bulk RNA-seq has reads from across transcripts and so quantification requires normalization by length. Having said that, if you are only interested in producing transcript compatibility counts then yes, this index can be used.