help WGS mixed cells

Question

[crakenmar]

I want to sequence whole genome from a mixed cells (mouse-human) in the same sample, is there any bioinformatic way to discriminate between both sequences and have parallel results, i mean, ibtain the WGS for human and for mouse separetly

Thank you

[crakenmar]

Thank you very much Ryan, the only problem I have is the tumor I am studying is human in vitro in mouse and infiltrates into the mouse brain, so I'm not sure when I pick up tissue if it would be contaminated with mouse genome.

Thanks a lot for your help

[Devon Ryan]

Yeah, have a look at the paper I linked to, from a collaborating group of ours. They routinely do effectively what you want, just with different cell-types from the same organism.

[crakenmar]

Thanks a lot :)

Answer 1

[Devon Ryan]

You're probably better off separating things prior to sequencing. The general idea is to simply tweak something like BiTS-ChIP to label according to species rather than cell-type. In short, you extract nuclei, label with a fluorescent antibody specific to a marker that can distinguish the two species (I'm sure there's something), and then flow sort before extracting DNA. If you can find a good antibody or two then this shouldn't be too difficult.

The only informatic way to proceed with a mixed cell population involves aligning to both and then assigning reads to the species to which they better align. This is similar to the allele-specific pipelines, so you could just tweak one of those if needed (there might already be something written for this). This is, of course, far less than ideal, since large numbers of reads will be ambiguous (i.e., they map equally well/poorly to both or the difference is small enough that you can't make the species call with high reliability).