First, Evolutionary Trace Annotation, or ETA, identifies which proteins have local evolutionary and structural features in common; next, these proteins are linked together into a proteomic network of ETA similarities; then, starting from proteins with known functions, competing functional labels diffuse link-by-link over the entire network.
Next, we obtained gene-function associations and the ontology of functional labels from the GO Consortium (Ashburner et al., 2000).
Since protein functions are inter-correlated and most functional labels often have a relatively small number of member proteins, these algorithms ignore the interrelationship among labels, which can often be used to boost the prediction accuracy [ 3, 19].
Results: We propose a novel function prediction algorithm, clusDCA, which transfers information between similar functional labels to alleviate the overfitting problem for sparsely annotated functions.
In addition, MNet takes into account the unbalanced label problem in protein function prediction, and incorporates a label weighted scheme into the unified objective function to give more emphasis to the functional labels with fewer proteins.
Then we consider the unique combinations of keywords as the protein functional labels (FL), which characterize the biological functions of the given protein and construct the contingency tables and graphs providing the projections of transcription units (TU) and alternative splice-variants (SV) onto all FL of the proteome of a given organism.
Once this higher level class structure is modelled however, one may employ it to infer added functions for known genes or, more interestingly, assign functional labels to unclassified ORFs.
After obtaining the low-dimensional vector representations of both genes and functional labels, we use these vectors to predict gene function.
Because many molecular functions (MFs) are inherently specific in their scope, a large number of functional labels have only a few annotated genes (or positive annotations); for instance, in the human GO annotation database (Ashburner et al., 2000), there are currently 8626 GO labels with at least 3 annotations, 4178 of which have <10 annotated genes and 7905 labels have <100 genes.
Because modeling the complex relationship between genes and functional labels with a single transformation matrix may be overly restrictive, we group functions into different clusters and learn a separate projection model for each cluster.
The fewer functional labels we have to compartmentalize what we do, and the more porous the boundaries between forms, the more likely we will be to resist ideas that don't seem to fit.
