The TransferAnnotations.jar method queries the network using the source sequences with functional annotations, one at a time. The functional annotations are tranfered from the source nodes to their homologs identified by RandomWalkwithRestart.jar, called target sequences and scores are assigned to the annotation taking into account the quality of the source annotations (user-defined), the number of homologs of the target sequences that have the annotation and the orthofuzz score between the souce and the target sequence.

Figure : Annotation flow network: A1-A3 are annotations of source sequences S1, S2 and S3. axsy represents the user-defined confidence of the annotation ax to be associated with sequence sy. sxty is the maximum score normalized orthofuzz score(sx,ty) obtained by querying the pairwise sequence similarity network using sx.

As shown in the figure X, the annotation scoring and assignment can be defined as a problem of multiplying the source annotation matrix with the orthofuzz matrix as shown in the figure X.

Figure 1: TransferAnnotation algorithm:

Link not working Link not working

Random Walk Clustering on the network starting from each source node with annotation. The format of the AnnotationGraph can be found here.

Figure 2: TransferAnnotation protocol:

Link not working

Downloads:

Check here for parameters of the program

Look here for an example run

Download the jar file here