λ\x | 4 | 5 | 6 | 7 | 8 |
---|
0.5 | 473/54.8% | 552/65.0% | 561/68.3% | 561/69.0% | 561/69.1% |
1.0 | 4,598/28.6% | 6,131/37.9% | 6,450/43.5% | 6,501/45.9% | 6,508/46.7% |
1.5 | 14,119/14.9% | 21,179/21.4% | 23,386/26.8% | 23,915/30.3% | 24,021/32.0% |
2.0 | 27,067/7.8% | 45,111/11.8% | 52,630/16.0% | 55,036/19.5% | 55,675/21.8% |
2.5 | 40,080/4.1% | 73,481/6.5% | 90,877/9.4% | 97,835/12.2% | 100,145/14.6% |
- The values are given as a function of fold coverage (λ, row labels) and the upper bound on the number of overlapping reads (x, column labels). For instance, at λ = 1.0 and x = 5, there are 6,131 correct SNP calls and 37.9% as many duplication-induced erroneous ones. This is a theoretical analysis based on informally fitting a model (see text) to data from the genome of Dr. James Watson.