Table 4

Assessment of the "Top 50 genes" page.


# Gene-Disease Associations in GAD
# correct Gene-Disease Associations not in GAD
# False Recognition
Precision of Gene Recognition
Examples False Recognition

RANK based method
35
9
6
88% (44/50)
IDDM2 (insulin dependent diabetes mellitus type II)
CD8A (CD8-Cells)
IgAN (IgA-Nephropathy)
ACR (Albumin-To-Creatinine-Ratio)
Frequency based method
34
11
5
90% (45/50)
IDDM2 (insulin dependent diabetes mellitus type II)
CD8A (CD8-Cells)
PLF (Pulmonary function)

The top 50 genes from the AIDB were evaluated in comparison to GAD (# genes found in GAD, correct genes not found in GAD and falsely recognised genes). The resulting precision of gene recognition is 88% for the RANK based method and 90% for the frequency based method. Erroneously recognised genes are listed in the last column.

Karopka et al. BMC Bioinformatics 2006 7:325   doi:10.1186/1471-2105-7-325