The Scientific World Journal

Research Article

Biomarker Identification for Prostate Cancer and Lymph Node Metastasis from Microarray Data and Protein Interaction Network Using Gene Prioritization Method

Algorithm 1

Overview of NetWalk Phase.
 Input: Differential Expression PPI DE_PIN, Pearson Correlation Coefficiente Networks (Control and
    Case) D_PCC_PIN and N_PCC_PIN, Seed Set as Training set of the RWR   𝜌 SeedSet, Boundaries
    and Step of Filtering ThresholdStart, ThresholdEnd, ThresholdStep
 output: Set of filtered networks DE_FilteredPPI
( 1 ) Create Stochastic Matrix StochasticMatrix from Case Samples Pearson Correlation Coefficient 𝑃 𝐶 𝐶 𝐷
 PIN according to  (2)
( 2 ) Create Reference Stochastic Matrix StochasticMatrixRef from Control Samples Pearson Correlation Coefficient 𝑃 𝐶 𝐶 𝑁 PIN according to  (2)
( 3 ) Run RWR with 𝛾 = 0 . 3 using StochasticMatrix save the results in RWR_Scores
( 4 ) Run RWR with 𝛾 = 0 . 3 using StochasticMatrixRef save the results in RWR_RefScores
( 5 ) Compute raw edge flux value for case sample e f 𝑖 𝑗 according to  (3)
( 6 ) Compute raw edge flux value for control (reference) sample: e f r e f 𝑖 𝑗 according to  (3)
( 7 ) Compute Normalized Edge Flux according to  (4)
( 8 ) for   𝑖 ← ThresholdStart to ThresholdEnd  step ThresholdStep  do
( 9 ) Create Edge List EdgeList using threshold + 𝑖 ( 𝑖 → + ∞ )
( 1 0 )   Change Weights in EdgeList to weights in DE_PIN using only the edges in EdgeList store results in DE_FilteredPPl(+i)
( 1 1 )   Create Edge List using EdgeList threshold − 𝑖 ( − ∞ → − 𝑖 )
( 1 2 )   Change Weights in EdgeList to weights in DE_PIN using only the edges in EdgeList store results in DE_FilteredPPl(−i)
( 1 3 )   Create Edge List EdgeList using threshold C 𝑖 ( − ∞ → 𝑖 ∪ 𝑖 → ∞ )
( 1 4 )   Change Weights in EdgeList to weights in DE_PIN using only the edges in EdgeList store results in DE_FilteredPPI(Ci)
( 1 5 )   return  Set of Filtered PPI  Networks DE_FilteredPPl