The number of both CNVs and de novo SNP changes that are reported are difficult to believe compared to the literature (e.g. A map of human genome variation from population-scale sequencing. 1000 Genomes Project Consortium, Durbin RM, Abecasis GR, Altshuler DL, Auton A, Brooks LD, Durbin RM, Gibbs RA, Hurles ME, McVean GA.
Nature. 2010 Oct 28;467(7319):1061-73).

There are many questions about the data analysis that were not clear from this paper. I have directly contacted the authors and the editors of Plos One with no response.

My main questions are:

1. Did you visually inspect each of the CNV regions (logR, b allele ratio) to determine whether differences between individuals were due to low sensitivity of CNV calling algorithms?

2. Usually when researchers use the SNPs from the Affymetrix SNP 6.0, ~200K SNPs are excluded due to low quality, e.g. low call rate, deviation from HWE, poor clustering. Did you use any external dataset to allow exclusion of such poor-quality SNPs (e.g. PLoS Genet. 2010 Oct 28;6(10):e1001183.). However it appears that no such QC criteria were applied here.

3. Did you inspect the cluster plots (i.e. A vs B allele intensity) for those SNPs where there appeared to be a genotype inconsistency within the family?

4. Are the raw CEL files available through GEO?