Table 1.
Genotyping accuracy on a subset of the HapMap CEU population
| cnvHiTSeq | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| cnvHiTSeq* | cnvHiTSeq† | |||||||||
| Location | Predicted Location | Predicted length (bp) | cnvHap r2 | SCIMM r2 | r2 | Accuracy | Missing rate | r2 | Accuracy | Missing rate |
| chr1:35098051-35115368 | chr1:35100671-35112111 | 11,440 | 1.00 | 1.00 | 0.77 | 0.88 | 0.23 | 1.00 | 1.00 | 0.00 |
| chr1:152759872-152770356 | chr1:152760173-152770753 | 10,580 | 0.90 | 0.94 | 0.85 | 0.82 | 0.23 | 1.00 | 1.00 | 0.00 |
| chr3:151625213-151657165 | N/A | N/A | 0.00 | N/A | N/A | N/A | N/A | N/A | N/A | N/A |
| chr7:97395305-97402641 | chr7:97395365-97402646 | 7,281 | 1.00 | 1.00 | 0.66 | 0.81 | 0.05 | 1.00 | 1.00 | 0.00 |
| chr7:115930472-115941073 | chr7:115931453-115941632 | 10,179 | 1.00 | 1.00 | N/A | 0.95 | 0.00 | 1.00 | 1.00 | 0.00 |
| chr8:51030941-51038331 | chr8:51031082-51038282 | 7,200 | 0.92 | 0.93 | 0.63 | 0.94 | 0.11 | 1.00 | 1.00 | 0.00 |
| chr8:144700485-144714694 | chr8:144700505-144714606 | 14,101 | 1.00 | 0.97 | 0.54 | 0.67 | 0.00 | 1.00 | 1.00 | 0.00 |
| chr10:71280989-71291079 | chr10:71280949-71291070 | 10,121 | 0.90 | 0.82 | 0.58 | 0.86 | 0.05 | 0.89 | 0.95 | 0.00 |
| chr11:5783630-5809284 | chr11:5784450-5809211 | 24,761 | 1.00 | 1.00 | 0.61 | 0.83 | 0.00 | 0.94 | 0.94 | 0.00 |
| chr11:107238222-107244154 | chr11:107238422-107244103 | 5,681 | 0.94 | 0.97 | 0.83 | 0.90 | 0.00 | 1.00 | 1.00 | 0.00 |
| chr15:34694542-34817215 | chr15:34701483-34817043 | 115,560 | 0.80 | 1.00 | 0.69 | 1.00 | 0.05 | 1.00 | 1.00 | 0.00 |
| chr15:76884597-76907042 | chr15:76884597-76896918 | 12,321 | 0.56 | N/A | 0.96 | 0.95 | 0.00 | 1.00 | 1.00 | 0.00 |
| chr19:35851153-35861684 | chr19:35851134-35863213 | 12,079 | 1.00 | 1.00 | 0.81 | 0.90 | 0.05 | 1.00 | 1.00 | 0.00 |
| chr19:52132525-52148984 | chr19:52132606-52149186 | 16,580 | 0.96 | 0.96 | 0.88 | 0.86 | 0.00 | 1.00 | 1.00 | 0.00 |
| chr20:1558407-1585809 | chr20:1561187-1585928 | 24,741 | 0.90 | N/A | 0.95 | 0.94 | 0.00 | 1.00 | 1.00 | 0.00 |
| chr22:23154417-23243496 | chr22:23186037-23241798 | 55,761 | 0.22 | N/A | 0.61 | 0.73 | 0.00 | 0.77 | 0.80 | 0.09 |
| chr22:24323894-24418396 | chr22:24343395-24397295 | 53,900 | 0.00 | N/A | 0.91 | 0.86 | 0.00 | 1.00 | 1.00 | 0.00 |
| chr22:39366812-39386139 | chr22:39358773-39383652 | 24,879 | 1.00 | 0.96 | 0.46 | 0.85 | 0.00 | 1.00 | 1.00 | 0.00 |
Genotyping accuracy as measured by the concordance between copy number estimates on 22 HapMap CEU samples from the low-coverage pilot of the 1000 Genomes Project and reference copy number estimates obtained using PCR. Concordance is quantified using two different metrics: the correlation coefficient r2 between the reference and the predicted genotypes as well as the fraction of calls with the correct genotype for both alleles. r2 measurements for SCIMM (SNP-Conditional Mixture Modeling) were obtained from the supplementary material of [16]. r2 measurements for cnvHap were obtained from the supplementary material of [9]. Two different versions of cnvHiTSeq were used: cnvHiTSeq*, which is a single-sample version of the algorithm that does not take advantage of the population modeling capabilities, and cnvHiTSeq†, which trains the parameters of the model using the entire low-coverage HapMap CEU population from the 1000 Genomes Project (currently consisting of 94 samples). The genotyping accuracy was calculated using 22 of the 94 samples, since these were the only samples for which PCR copy number estimates were available. When all the samples are predicted to be copy neutral for a given location, the accuracy and r2 are undefined and denoted by N/A. cnvHiTSeq calls with posterior probabilities lower than 80% were excluded and declared as missing.