Closed jliu678 closed 3 years ago
Thank your for the bug report! And thanks for including the data and the traceback. This is likely a duplicate of https://github.com/ChristophH/sctransform/issues/71.
Could you try the current version of sctransform in the develop branch?
remotes::install_github("ChristophH/sctransform@develop")
Thank your for the bug report! And thanks for including the data and the traceback. This is likely a duplicate of #71.
Could you try the current version of sctransform in the develop branch?
remotes::install_github("ChristophH/sctransform@develop")
still not working... attached is the traceback and session info. it looks the same as previously except attached "sctransform" was development version now.
By the way, I think it is about the update made in the newest version compared with version 0.3, which running well.
You are setting min_cells = 1
which means that sctransform is keeping all genes that are detected (non-zero count) in at least one cell. We do not advice doing so, and here it breaks the parameter estimation.
I would argue that genes that are detected in just one cell are not going to have a large influence on the downstream analysis and can safely be dropped. In your case, setting min_cells = 2
or higher seems to be sufficient for the parameter estimation to pass.
The default value for min_cells
is 5, and we should probably issue a warning when this parameter is set lower than that.
Hi there,
"bugmtx.rds" containing the matrix "bugmtx" and the txt file that prints the "err trackback" have been attached in the zip file,and
S <- CreateSeuratObject(counts = bugmtx, min.features = 1,min.cells = 1)
S<- subset(x = S, subset = nFeature_RNA > 200)
S<- SCTransform(S, verbose = T,return.only.var.genes=F, min_cells = 1)
gives err message in console:
Calculating cell attributes from input UMI matrix: log_umi
Variance stabilizing transformation of count matrix of size 11743 by 255
Model formula is y ~ log_umi
Get Negative Binomial regression parameters per gene
Using 2000 genes, 255 cells
|===================== | 25%
warning: solve(): system seems singular; attempting approx solution
Error in h(simpleError(msg, call)) :
error in evaluating the argument 'x' in selecting a method for function 't': missing value where TRUE/FALSE needed
In addition: There were 50 or more warnings (use warnings() to see the first 50)
Called from: signalConditions(obj, exclude = getOption("future.relay.immediate",
"immediateCondition"), resignal = resignal, ...)
session info:
hope to hear your advice soon
thanks
data file and err traceback.zip