missing value?

[yingzhang121]

Hi,

Is there an option in determineDropoutCandidates to handle missing value?

I made a reduced UMI count matrix for my single-cell RNA-Seq data. Then I had the "missing value" issue:

x=read.csv("filtered_reduced_matrix.csv", header=T, row.names=1, skipNul=T)
scPan <- scDataConstructor(as.matrix(x))
scPan <- determineDropoutCandidates(scPan)
Error in density.default(object@nData[, topLibraries[i]], kernel = "epanechnikov", :
'x' contains missing values

Best,
Ying

[hojwk]

Hi Ying,

Sorry it has taken some time to get back to you.

At the moment, detemineDropoutCandidates does not allow for missing values (e.g., NA). The call to the function "density" from the R "stats" package, implicitly has the parameter "na.rm" set to FALSE - which will produce an error when encountering any missing values. If we explicitly set na.rm to TRUE in our CIDR code, this would "remove any missing values from the data", which we do not think is desirable. If a missing value indicates a possible dropout, we recommend replacing all occurrence of NA to zero. I hope this is helpful.

Regards,
Joshua

[yingzhang121]

Thank you for your comments.

I managed to run CIDR on my reduced matrix by running the following command before the scPan commands:

x<-na.omit(x)

Do you see any caveat in it?