TABLE 1.

Gene Regulatory Network Inference Algorithm Using Random Walk with Restart.

Algorithm: RWRNET
Input: Gene microarrays data G = {g1,⋯,gN}
Output: A gene regulatory network
1: Construct a MI matrix MI according to Eq. 1;
2: Calculate restart probability α using Eq. 8;
3: Construct transition probability matrix W using Eq. 11;
4: Calculate gene expression level EL(gi) for each gene using Eq. 9;
5: Select centres of functional module and put them into set C according to Eq. 10;
6: Construct functional modules: moduleg1 = {g1}, moduleg2 = {g2}, ⋯, modulegN = {gN};
7: For each gene gi ∈ C do
8: Rank the genes gj in {G−gi} according to MI(gi, gj) in descending order to form ranking list MIL;
9: modulegi ← the top log⁡N genes in MIL;
10: End For
11: For each gene gi ∈ G do
12: Construct initial probability vector $p_0^{\left(g_i\right)}$ according to modulegi;
13: $p_{t+1}^{\left(g_i\right)}=\mathit{RWR}(\mathrm{\alpha },W,p_0^{\left(g_i\right)})$;
14: Calculate final score MIP(gi) according to Eq. 13;
15: End For
16: Infer network using Eq. 14;
17: Process isolated genes based on IPC-MB;
18: Return the optimised gene regulatory network.