we see no clear evidence of this in comparison of the HapMap and non-HapMap samples in our study. Not all experiments passed quality control thresholds, but this was mainly the case for the lower resolution platforms (Affymetrix 250K, Illumina 650Y). The quality control steps also highlighted the problems of using an external reference set for analysis of Affymetrix data. Both for the 250K array run at TCAG and the Affymetrix 6.0 arrays run at WTSI, the lack of an internal reference led to poor signal-to-noise ratios. Finally, the different laboratories involved may be more experienced processing certain array types, leading to relatively better results, and the data inevitably contain subtle batch effects across sites and time points that is present in all data sets49. Nevertheless, we believe that the current study is representative of the results being obtained at different laboratories processing these arrays.
