Table 1. Summary statistics from two experiments (A and B), each of which examined copy number in the same 12 DNA samples using 40 probes.
| Probe type | Sample | Experiment | Mean ± SD | No. of bands scored |
|---|---|---|---|---|
| Autosomal | Control | A | 1 ± 0.10 | 330 |
| B | 1 ± 0.13 | 342 | ||
| X-linked (female) | Control | A | 1.00 ± 0.10 | 48 |
| B | 1.01 ± 0.11 | 43 | ||
| X-linked (male) | Control | A | 0.499 ± 0.045 | 69 |
| B | 0.496 ± 0.072 | 62 | ||
| Combined data | ||||
| All ‘diploid’ (autosomal + X-linked female) | (A+B) | 1 ± 0.114 | 763 | |
| All ‘haploid’ (X-linked male + deletions) | (A+B) | 0.497 ± 0.062 | 147 | |
| Inter-assay variation | A versus B | 0.060 (pairwise SD) | (468) |
| Incorrect calls (using a threshold ratio of 0.75) | Predicted | Observed | ||
|---|---|---|---|---|
| False positive | 1.4% | 0.9% (7/763) | ||
| False negative | 0.03% | 0% (0/147) |
The combined data pooled from both experiments is used to derive normal distributions for copy numbers of 2 (‘diploid’) and 1 (‘haploid’), and hence to predict the frequencies of incorrect calls. The inter-assay variation shows the mean pairwise standard deviation from 468 pairwise comparisons of corresponding results from the two experiments.