[BioC] | Correlation | > 1 (limma)

[Gordon Smyth]

Hi Naomi,

>[BioC] | Correlation | > 1 (limma)
>Naomi Altman naomi at stat.psu.edu
>Sat Nov 12 20:54:37 CET 2005
>
>I just used intraspotCorrelation as suggested in the online Limma
>documentation for single channel analysis.
>Over half of my estimated correlation coefficients are bigger than
>1.0 in absolute value.
>
>Naomi S. Altman                                814-865-3791 (voice)
>Associate Professor
>Dept. of Statistics                              814-863-7114 (fax)
>Penn State University                         814-865-1348 (Statistics)
>University Park, PA 16802-2111

Just to make it confusing :) the correlations are stored on the atanh 
(unconstrained) scale. To get correlations on the usual scale,

 > corcoef <- tanh( corfit at all.correlations )

Cheers
Gordon