| Analysis | To select parameters for calling CNVs (that is, the statistical threshold of the ADM2 algorithm, the minimum +/- log2 ratio and the minimum number of consecutive probes in a CNV interval), we calculated the sensitivity and positive predictive value based on the comparison of aCGH-based CNV calls (using our high-resolution Agilent 24M platform) with read-depth sequence data for two samples from Korean individuals (AK1 and AK2). We attempted to obtain `absolute' copy number status of the sample from NA10851, which was used as the reference sample for aCGH experiments in this study. For this, we used read-depth data for NA10851 obtained from massively parallel sequencing by the Illumina GA II data. The read-depth data represent the copy number status of NA10851 as compared to the human reference genome (hg18) because the short read sequences were aligned to hg18. |
