Fig. 7
From: ModularBoost: an efficient network inference algorithm based on module decomposition
The workflow of ModularBoost. (a) Input: time-stamped single-cell gene expression data; (b) Step 1: based on expression patterns, ICA-FDR assigns the genes into several modules with various colors; (c) Step 2: GRNBoost2 infers GRN for each gene module separately and sorts \(n\_comps\) groups of scores in descending order; (d) Step 3: the inter-modular interactions are computed by sparse regression; (e) Step 4: normalization of inference scores separately, and computation the amalgamated edge predictions of the GRN