Figure 1

Scheme of BRIDGE. (A) A multiple linear regression model is proposed to explain the phenotypic similarity between two diseases using functional similarities between the two sets of genes associated with the diseases. The regression function is $S_{d{d}^{\prime }}={\alpha }_d+{\displaystyle {\sum }_{i=1}^5{\beta }_d^i\left({\displaystyle {\sum }_{g\in G\left(d\right)}{\displaystyle {\sum }_{g^{\prime }\in G\left(d^{\prime }\right)}{S}_{g{g}^{\prime }}^i}}\right)}$, where S ⁱ _dd’ is the phenotypic similarity between two diseases d and d’, S ⁱ _gg’ the functional similarity between two genes g and g’ derived from the i-th data source, G(d) and G(d’) genes associated with diseases d and d’, respectively. We consider five genomic data sources (PPI, GS, GE, KEGG, and GO) in our model. (B) Given a query disease g and a candidate gene d, we assume the candidate gene is the only one associated with the disease, i.e. G(d) = {g}, and we calculate the coefficient of determination (R²) of the fitted model as a score to measure the strength of association between the disease and the gene. (C) Repeating (B) for every candidate gene, we obtain a score for each candidate. We then rank the candidate genes in non-increasing order according to their scores to obtain a ranking list.