Skip to main content

Table 8 Statistics about the predictions made by the base kNN classifiers and their label similarity-incorporated versions for the genes in Mnainmeh and Rosetta data sets not covered in the cross-validation experiments, and how many of these predictions could be validated based on the annotations added to the GO database between February–September 2008.

From: Incorporating functional inter-relationships into protein function prediction algorithms

Dataset

# Test genes

## Validated predictions

Median of ranks in label similarity kNN

Median of ranks in base kNN

P-value

Mnaimneh

1621

1003

497

637

0.0047

Rosetta

1609

998

509

560

0.319