Table 1.
Quality cutoff | Total number of SNPs | Correctly identified | False | positives | False | negatives | |
---|---|---|---|---|---|---|---|
A. SCN1A | |||||||
novoSNP | 10 | 447 | 33 | 414 | 92.6% | 0 | 0.0% |
15 | 122 | 32 | 90 | 73.8% | 1 | 3.0% | |
20 | 36 | 26 | 10 | 27.8% | 7 | 21.2% | |
25 | 26 | 22 | 4 | 15.4% | 11 | 33.3% | |
PolyPhred | 20 | 586 | 30 | 556 | 94.9% | 3 | 9.1% |
25 | 510 | 30 | 480 | 94.1% | 3 | 9.1% | |
50 | 347 | 30 | 317 | 91.4% | 3 | 9.1% | |
75 | 254 | 30 | 224 | 88.2% | 3 | 9.1% | |
95 | 208 | 30 | 178 | 85.6% | 3 | 9.1% | |
99 | 189 | 26 | 163 | 86.2% | 7 | 21.2% | |
PolyBayes | 0.1 | 54 | 18 | 36 | 66.7% | 15 | 45.5% |
0.25 | 46 | 17 | 29 | 63.0% | 16 | 48.5% | |
0.5 | 37 | 16 | 21 | 56.8% | 17 | 51.5% | |
0.75 | 33 | 16 | 17 | 51.5% | 17 | 51.5% | |
B. MAPT | |||||||
novoSNP | 5 | 3424 | 442 | 2982 | 87.1% | 10 | 2.2% |
10 | 1146 | 421 | 725 | 63.3% | 31 | 6.9% | |
15 | 484 | 377 | 107 | 22.1% | 75 | 16.6% | |
20 | 251 | 244 | 7 | 2.8% | 208 | 46.0% | |
25 | 206 | 203 | 3 | 1.5% | 249 | 55.1% | |
PolyPhred | 20 | 2637 | 280 | 2357 | 89.4% | 172 | 38.1% |
25 | 2510 | 280 | 2230 | 88.8% | 172 | 38.1% | |
50 | 2243 | 271 | 1972 | 87.9% | 181 | 40.0% | |
75 | 1892 | 252 | 1640 | 86.7% | 200 | 44.2% | |
95 | 1677 | 207 | 1470 | 87.7% | 245 | 54.2% | |
99 | 1572 | 175 | 1397 | 88.9% | 277 | 61.3% | |
PolyBayes | 0.1 | 991 | 140 | 851 | 85.9% | 312 | 69.0% |
0.25 | 830 | 136 | 694 | 83.6% | 316 | 69.9% | |
0.5 | 672 | 126 | 546 | 81.2% | 326 | 72.1% | |
0.75 | 567 | 115 | 452 | 79.7% | 337 | 74.6% |
For the SCN1A data set, the lowest novoSNP shown cutoff is 10 since all SNPs were found at this cutoff value.