Which region bulk DNA should I use?

[wyqhyrz]

For example, I have different time point scRNA sequencing data to explore the clonal structure of these samples. How do I choose the bulk DNA time point? And scRNA-seq and bulk DNA sampling from different regions, the bulk DNA may not show the accurate CNV or SNV of the scRNA-seq region's state, isn't it?

Looking forward to a reply from you. Thanks in advance for any help!

[junseonghwan]

I think Bulk DNA-seq should be performed at the later of the two time points.

For multi-regional study, it is possible that clonal evolution is occurring independently. You can probably check for that from the data. If they share common mutations for example. Once you can verify that they share common traits, you can then perform clonal analysis by combining them. In my opinion, when the regions are from a solid contiguous tumor mass, they are likely to derive from a common progenitor.

[wyqhyrz]

Sorry, I still don't get it. 1. In PhylEx, can we sequence several time points scRNA and only one time point DNA to construct the clonal state of the samples? Or do we sequence more time points of bulk DNA may be better? 2. If we just choose one-time point bulk DNA, why choose the later time point? 3. Maybe I didn't really understand the model of PhylEx, the formula is difficult for me to understand. I mean, the cells were sequenced by scRNA, they weren't able to be sequenced by bulk DNA, so the data of scRNA and bulk DNA are from different cells. Actually, I think I don't know the role of bulk DNA in PhylEx model. Thanks in advance for any help!

[junseonghwan]

1. Yes. There can be multiple time points for scRNA-seq and only one time point for bulk. If you have more time points of bulk DNA-seq, that should also work in theory by treating them as multi-regional bulk DNA-seq data. However, PhylEx wasn't designed and tested for that use case.

2. The later time point should contain the latest evidence of evolution. For example, if bulk sequencing is performed at time t0 and scRNA-seq at time t0 and at a later time t1, then there is risk that scRNA-seq data at time t1 contains mutations that are not found in the bulk. So the opportunity to learn about the evolution that took place since t0 is lost. By performing bulk DNA-seq at t1, it should contain mutations at t0 and t1 because mutations are inherited. If you can perform bulk at both t0 and t1, that would be even better simply because having more information can't hurt but as I said in point 1, PhylEx was not tested for that use case. It would be interesting to try. Do you have such data?

3. Yes, bulk and scRNA-seq are different cells and that is expected. We are trying to align the two modalities using the evidence (mutations) found in the cells to the evidence (mutations) found in the bulk. The underlying assumption is that evolutionary dynamics is shared across the cells sequenced using scRNA-seq and the cells in the bulk, which is a collection of large number of cells. From statistical sampling perspective, bulk represents a population level data while single cells represent more fine grained individual level data.

Please let me know if you would like further clarification.

[wyqhyrz]

Thanks for your reply! I am interested in PhylEx, but we haven't sequenced the samples. If I get the data(10X), I will try to apply PhylEx.

[junseonghwan]

No problem. The current version of PhylEx performs best on Smart-seq data but it would certainly be worth a try. We are also developing another method that will work better with 10x. Maybe it will be ready by the time you get the data. Good luck with your work!