[BioC] Illumina replicates - packages

Sean Davis sdavis2 at mail.nih.gov
Mon Oct 15 16:50:31 CEST 2007


Ingrid H. G. Østensen wrote:
> Hi
> 
> I am not sure this is the right place to post this message but maybe someone have some tips/experience to share.
> 
> We have an ongoing discussion regarding the number of replicates needed when we are using Illumina arrays. It is not easy saying one right number, but the main problem is this: If one can take advantage of the 30 something beads on the array can one use only one replica per sample? Have anyone done anything like this?

Hi, Ingrid.

The replicate beads allow an estimate of the technical variation _within
an array_ for the _same oligo probe_.  They do not allow estimation of
biologic variation or array-to-array variation.  If you do not need
estimates of the latter two types of variation, which is not a likely
circumstance, then you will need to do biologic replicates.  The
array-to-array variation on illumina, we have found, is quite low, but
it is still probably quite worthwhile to balance an experiment with
respect to the arrays whenever possible to reduce the possibility of a
technical issue producing what looks to be biological effect.

Sean



More information about the Bioconductor mailing list