comparing single cell RNA-seq and bulk RNA-seq results

Question

[saba.valadkhan]

Hi everyone, I have bulk cell RNA-seq data from a rare cell type, and would like to use the existing RNA-seq data in public libraries to confirm the identity of the cell. The only available RNA-seq study on this cell type is done using a single cell RNA-seq approach, and the study includes several other cell types from the same tissue too, so it's a very useful dataset. However, I'm not sure what the best approach is for comparing bulk cell and single cell RNA-seq results. The best idea I have come up with is randomly sub-sampling our bulk RNA-seq fastqs to generate several smaller fastqs each the approximate size of the individual single cell libraries in the target dataset. Is this a good idea, and is there a better way of doing this? The final goal is to get some unbiased grouping data, such as multidimensional scaling analysis, showing that our cells segregate with the right kind of cell in the single cell results. Thank you very much for your help!

[lhaiyan3]

Hi, Did you find the good way to do this? Thanks.

HY

Answer 1

[Chris Fields]

A talk recently by Cole Trapnell did something like this when assessing results for his tool Monocle (for single-cell RNA-Seq analysis). I did some digging, Fig. 1 in his pub assess gene expression levels in single-cell vs bulk RNA-Seq.

[saba.valadkhan]

Thank you very much for taking the time to search and find the article. I was only aware of the work of Marinov et al (2014, Genome res, Wold lab) comparing single cell and bulk RNA-seqs, so this added reference was very helpful. Despite the lower complexity of the single cell RNA-seqs, the overall gene expression patterns seem to correlate in dimensional reduction analyses in the assays I ran. Thanks again!