[BioC] pcaMethods and categorical NAs

Antoni Picornell urnenfelder at gmail.com
Thu Jan 14 11:52:57 CET 2010


Hello,

I'm trying to impute missing values in a matrix built with genotypes. The 
values in this matrix are the "usual" ones in this situation: 0,1,2 for 
homozygous(0 and 2) or heterozygous (1) situations. Thus, I have a matrix with 
0,1,2 and NAs.

I want t impute these NAs. I have found that the package pcaMethods could be a 
nice way to impute NAs in huge ammounts of data. The other solution I have 
found is fastPHASE, but it is very slow. However, as far as I know pcaMethods 
needs a numerical matrix.

Anybody knows whether this package can be applied in this situation? I have 
not seen anything related to this issue in the vignettes. Do you use another 
package to do this?

Thanks,
Antoni Picornell.



More information about the Bioconductor mailing list