Table 4

Proportion of misclassified samples. The table shows an detailed analysis of sample assignment to integrity categories. Each category ck is tested against a series of set union ∪ci of all other categories i with i = 1,...,k - 1. The RIN of each electropherogram is rounded up or down to the closest categorical value. The classification performance for assigning electropherograms to the two resulting sets is then measured by Receiver Operating Characteristics, more precisely by the area under the curve (cf. fig. 5). There are always a few interchanges between adjacent categories, but rarely interchanges over two categorical borders. Values are rounded to two decimals.

9

8

7

6

5

4

3

2

1


10

0.96

0.98

1.00

1.00

1.00

1.00

1.00

1.00

1.00

9

0.96

1.00

1.00

1.00

1.00

1.00

1.00

1.00

8

0.98

0.99

1.00

1.00

1.00

1.00

0.99

7

0.98

0.99

0.99

0.99

0.99

0.98

6

0.96

0.98

0.98

0.99

0.98

5

0.95

0.97

0.98

0.98

4

0.98

0.99

0.98

3

0.99

0.98

2

0.95


Schroeder et al. BMC Molecular Biology 2006 7:3   doi:10.1186/1471-2199-7-3

Open Data