Figure 1

Functional annotation decision rules. (A): Local network representation of annotated neighboring proteins (rectangles) with weighted links to the unannotated target (circle). Color coding indicates the pathways in which the annotated proteins participate. Two metrics can be considered when determining annotation predictions of a target protein: (1) occurrence frequency of an annotation in the neighbors (2) weights of the relevant links contributing to a particular annotation. MR employs occurrence-frequency metric alone and weights a candidate annotation by counting neighbors having that annotation; e.g. for the grey annotation, the score is 5*1 = 5. NW emploits both metrics and uses a weighted sum of the links for the relevant annotation; e. g., for the grey annotation, the score is 0.1 + 0.4 + 0.2 + 0.3*2 = 1.3. MW employs linkage-weight metric alone and weights the annotations by the maximal linkage weight among all the linkages contributing to a particular annotation; for the grey annotation, the score is Maxf = 0.4. Table (B) shows the annotation ranking lists in descending order.