Each validation run was validated based on standard curve and the following acceptance criteria for QCs: accuracy of 70 130% from nominal value, CV ≤ 25%, and at least 4/5 determination of each QC passing the criteria.
The "All" group thus contains samples with merged QC pass lanes, samples with merged QC fail lanes, and samples with merged QC pass and QC fail lanes.
Samples lacking two replicate lanes in either the QC pass or QC fail group were excluded from this analysis.
For the reproducibility comparison (Additional file 1), Pearson r was calculated using 2 replicate lanes corresponding to each sample represented in the QC pass and QC fail groups.
> -wrap-foot> Calls from Illuminus and GenoSNP produce the most discordant results at QC (with 5157 and 8815 unique QC passes and fails, respectively) whereas GenCall and optiCall appear to have more overlapping QC outcomes with other callers.
To assess clustering quality and accuracy, 600 unique QC pass SNPs and 600 unique QC fail SNPs were selected at random and manually called using a modified version of Evoker (Morris et al., 2010).
If QC is passed, the primary hits are selected for a second round of screening.
In the example given above, as the coverage of 1X, 4X, 10X and 20X supporting reads decreases slowly, this QC is passed.
Further details on genotyping and QC criteria (n = 5919 passed QC) can be found in a previously published study (33) (http://www.wtccc.org.uk/ccc2).uk/ccc2
To investigate whether the additional noise seen in QC failed samples impacted sequencing reproducibility, we computed the Pearson correlation between replicate lanes of samples that passed QC vs. failed QC (Additional file 1).
