[BioC] sample size estimation in genome wide association studies / packages

[Richard Pearson]

Hi Philippe

I'm certainly no expert on this, but you might want to look at the 
Wellcome Trust Case Control Consortium paper from last year:
http://www.nature.com/nature/journal/v447/n7145/abs/nature05911.html

They have a discussion of sample size (see, e.g. Figure 6), and I 
believe concluded that without the thousands of samples they used, they 
wouldn't have detected many of the associations they did find. So yes, 
50 could really be way too few.

Best regards

Richard.


phguardiol at aol.com wrote:
> Dear Bioconductor users,
>
> I would like to get "the feeling" of experts dealing with SNP microarrays regarding sample size estimation in genome wide association studies.
> Some of the physicians / doctors who would like to perform such studies are asking me about sample size estimation, that is to say?the minimal?number of?chips to be run so that the analysis will?bring... let say positive results.
>
> My point of view is that in these studies we are more likely to generate hypothesis than to test hypothesis, so that the question of sample size?does not seem to be appropriate for me. 
> However,?running 50 chips for this kind of study could be really too short. 
>
> I have read again all the papers published in the new england journal of medicine in 2007 and I have not found anything about this question of sample size... or minimal number of chips.
>
> Point of view of experts would be greatly appreciate. 
>
> If this issue is "real" which package would you recommend for this kind of estimation ??
>
> Thanks
>
>
> Philippe?G
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
>
>