search

meichendong / SCDC

SCDC
42 stars 9 forks source link

Too few common Genes #1

Open Zanvine-gitcode opened 4 years ago

Zanvine-gitcode commented 4 years ago

Hello,

Firstly, thanks a lot for your tool, So far I've found your vignette really intuitive and your package really easy to use.

I'm currently testing how well deconvolution works on 'ground-truth' bulk samples taken from GTEx (i.e. their cultured fibroblast cells).

I get this error when trying to estimate the proportions using SDCD_prop() & SCDC_ENSEMBLE()

cts <- c("Epithelial","Tcell","NK-cell","Macrophage", "Fibroblast","Endothelial") ## Cell-types present in my sc2.es object

res <- SCDC_prop(bulk.eset = Fib.es , sc.eset = sc2.qc$sc.eset.qc, ct.varname = "Cell-Type", sample = "SampleID", ct.sub = cts, truep = NULL )

Creating Basis Matrix adjusted for maximal variance weight Error in SCDC_prop(bulk.eset = bulk.eset, sc.eset = zz, ct.varname = ct.varname, : Too few common genes!

min(nrow(Fib.es), nrow(sc2.qc$sc.eset.qc)) [1] 21579 ## Min number of genes present in the ExpressionSets are 21579

length(intersect(rownames(Fib.es),rownames(sc2.qc$sc.eset.qc))) [1] 4294 ## 4294 common genes.

Any advise on parameters or methods I can use to address this issue?

Many thanks in advance.

meichendong commented 4 years ago

Hi! Thanks for using SCDC! I don't see this issue from my end, so if you could provide a toy data I can help to figure this out!