mz2 / peer

A factor analysis package
9 stars 41 forks source link

PEER residuals 'NA' with 45 hidden covariates #7

Open gouthamatla opened 5 years ago

gouthamatla commented 5 years ago

Hi,

I am running PEER on 292 samples with 14000 genes. I set the number of hidden variables to find to 45. It returns a matrix of residuals with all values set to NA. I used all default parameters except the number of hidden factors.

It was ok when I set the hidden factors to 15 but failing on 45. May I know the reason ? Should I increase the iterations ? I do not see any sign of convergence with 15 or 45.

Thanks, Goutham A