gene expression correlation positive controls scRNAseq

Question

[guillaume.rbt]

Hi all,

I'm currently testing tools to assess the correlation of gene expression with scRNAseq human data.

I would like to have housekeeping "positive controls" of two genes that are always highly correlated, in any type of cell.

Would anyone know an example of a pair of gene like this?

Thanks

[igor]

One potential resource: https://www.tau.ac.il/~elieis/HKG/

Of course, with the dropouts in scRNA-seq, it's going to be difficult to get any meaningful correlations.

[guillaume.rbt]

thanks ! I've found ABCF1 and ABCF2 that seems to be good candidates (R=0.81 on all TCGA normal samples)

[igor]

TCGA is not single-cell. Did you mean "RNAseq" instead of "scRNAseq" in the title?

[guillaume.rbt]

I used TCGA to check quickly the correlation between the genes, but I indeed work on scRNAseq.

Answer 1

[GouthamAtla]

In this paper, Benchmarking algorithms for gene regulatory network inference from single-cell transcriptomic data, Nat Methods, authors did generate loads of simulated data sets as ground truth. Their method allows to plugin any algorithms and let you asses your method.

You can access to their synthetic data and code on GitHub .

BTW, its already a benchmarking paper. Might save time for you in testing methods.