From: Prioritizing disease genes with an improved dual label propagation framework
NOTATION | DESCRIPTION |
---|---|
n | Number of genes |
m | Number of phenotypes |
X i∙ | i-th row of matrix X |
X ∙j | j-th column of matrix X |
\(\boldsymbol {W}_{1}\in \mathbb {R}^{n\times n}\) | Binary PPI network |
\(\boldsymbol {W}_{2}\in \mathbb {R}^{m\times m}\) | Phenotype similarity network |
\(\bar {\boldsymbol {S}}_{1}\in \mathbb {R}^{n\times n}\) | Normalized PPI network |
 | \(\bar {\boldsymbol {S}}_{1}=\boldsymbol {D}_{1}^{-\frac {1}{2}}\boldsymbol {W}_{1}\boldsymbol {D}_{1}^{-\frac {1}{2}}\) |
\(\bar {\boldsymbol {S}}_{2}\in \mathbb {R}^{m\times m}\) | Normalized phenotype similarity network |
 | \(\bar {\boldsymbol {S}}_{2}=\boldsymbol {D}_{2}^{-\frac {1}{2}}\boldsymbol {W}_{2}\boldsymbol {D}_{2}^{-\frac {1}{2}}\) |
\(\hat {\boldsymbol {Y}}\in \mathbb {R}^{n\times m}\) | Known binary gene-phenotype associations for training |
\(\boldsymbol {{Y}}\in \mathbb {R}^{n\times m}\) | Gene-phenotype associations matrix to be learnt |
\({\boldsymbol {S}}_{1}\in \mathbb {R}^{n\times n}\) | Weighted PPI network to be learnt |
\({\boldsymbol {S}}_{2}\in \mathbb {R}^{m\times m}\) | Weighted phenotype similarity network to be learnt |