Skip to content
Prev 46388 / 63424 Next

Problem with R >3.0.0

Hi Sam,

I assume you mean that correlation for _genes_ (not samples)
11262:30000 is 0? I am the maintainer of the WGCNA package but
unfortunately I don't have access to a Mac big enough to try
30000x30000 correlation matrix, but I would be thankful if you could
try reproducing the problem with smaller matrices (e.g. 20000x20000)
and try to produce a small reproducible example by
generating the data using say rnorm, say like this:

nGenes = 20000 # as small as possible that still produces the error
nSamples = 100
datExpr1 = matrix(rnorm(nSamples * nGenes), nSamples, nGenes)

simMat = bicor(datExpr1, use = 'p')

Best,

Peter

On Mon, Aug 19, 2013 at 11:53 AM, Shelton, Samuel
<SheltonS at stemcell.ucsf.edu> wrote: