## Table 4 |
||||||

Summary of coverage, total number of candidate variants (all polymorphisms) and number
variants called by the variant detection model in Q30 ORP for the 3 natural viral
samples |
||||||

Natural sample |
Mean ORP coverage |
# candidate variant calls at Q≥30 |
# variants called at error rate 5×10^{-5} |
# variants called at error rate 5×10^{-4} |
FDR for error rate 5×10^{-5} |
FDR for error rate 5×10^{-4} |

BCV | 79,037x | 20413 | 2133 | 152 | 19% | 0% |

Fox1 | 80,546x | 14472 | 1354 | 70 | 21% | 0% |

Fox2 | 96,198x | 20417 | 1596 | 88 | 26% | 0% |

The number of candidate variants served as the Bonferroni correction factors in the
variant detection model. The number of false discoveries in each sample at error rate
5×10^{-5} is estimated by taking 2% (coverage adjusted false positive rate derived from control
plasmids, Additional file 1: Table S2, ) of the number of candidate variants. The false discovery rate (FDR)
is the proportion of false discoveries among variants discovered, e.g. for BCV, it’s
the ratio between 2% of 20413 and 2133, i.e.19%. FDR at 5×10^{-4} is 0% for all samples since the estimated false positive rate at this error rate
is 0% (Additional file 1: Table S2).

Chen-Harris * et al.*

Chen-Harris * et al.* *BMC Genomics* 2013 **14**:96 doi:10.1186/1471-2164-14-96