Hi all,

When running svaseq as "batch_sup_sva = svaseq(dat0,mod1,mod0,controls=controls,n.sv=1)$sv", I got an error:
Error in svd((dats - rowMeans(dats))) : a dimension is zero

Does anyone know how this problem was happened?

Many thanks,
Yi

Hi Yi, 

This could be happening for a couple of reasons: (1) you have a row or column that is all identical or all equal to zero, (2) because the controls vector is all FALSE or has only one TRUE in it because then you'd only have a vector not a matrix, or (3) you have a bunch of NAs in the matrix. If you could check and see that the data you are passing (looks like dat0) has two dimensions both greater than 1, that the controls vector has more than one TRUE, and that you have no constant or near constant rows or columns that would help diagnose the problem. 

Cheers

Jeff