We have extended the supervised framework of Yeung et al. [ 3] in three ways.
Third, we refine the supervised framework to adjust for sampling bias towards positive cases in the training data, thereby calibrating the prior distribution.
The notion of relevance can be cast into a supervised framework by considering that for each one of the x ij features belonging to the feature subset, the relevance function ρ is defined as follows [1]: ρ : R F × T × K → R F × T ( X, C, x ij ) ↦ ρ ( X, C, x ij ) ∈ R + (3).
To account for external knowledge in the network construction process, Yeung et al. [ 3] introduced a supervised framework to estimate the weights of various types of evidence of transcriptional regulation and subsequently derived top candidate regulators.
Ensembles of classifiers that utilize bagging, boosting, and hybrid-approaches for imbalanced datasets in the supervised framework were reviewed by Galar et al. [ 13].
The supervised framework is similar to the supervised PCA (SPCA) method first proposed for pathway-based gene expression analysis and GWAS based on traditional PCA [ 19, 25].
Additional file 2: Table S3 gives the estimated regression coefficient and the posterior probability for each external data type in our revised supervised framework.
For each gene g, rank the candidate regulators based on the regulatory potentials predicted from the supervised framework.
We have derived CPCA for aggregated association analysis of categorical SNP data, which is further extended to SCPCA in a supervised framework.
A supervised framework was used to estimate the relative contribution of each type of external knowledge and from this a shortlist of promising regulators for each gene was predicted.
While the supervised framework was shown to enrich current PPI data with additional inferred PPIs, its applicability is still limited.
