The analyses are implemented in R and Julia programming languages.

Please find the relevant files in SyntheticDataAnalysis/FullNetwork_SampledNetwork_Comparison folder.

1. Run PrepareAttractorSet.ipynb file (with R kernel). GA library is required. Configurable parameters are n_genes= number of genes, num_sub_bn_sets=number of samples per attractor set for building sampled networks, num_sel_bits=number of selected nodes for building sampled networks, num_bits=number of nodes.

2. Run full_network_constructBNs.sh from command line. Configurable parameters are Attractors_folder=The folder where the BN construction algorithm will find attractors, output_folder=The folder where results from BN construction algorithm be saved, num_bns= Number of times the BN construction algorithm will run, n_sets=Number of attractor sets.

3. Run sampled_network_constructBNs.sh from command line. Configurable parameters are Attractors_folder=The folder where the BN construction algorithm will find attractors, output_folder=The folder where results from BN construction algorithm be saved, num_bns= Number of times the BN construction algorithm will run, n_sets=Number of attractor sets

4. Run PrepareDataForPlots.jl from command line. DelimitedFiles,DataFrames, FileIO, JLD2,Statistics are the required packages.
   Configurable parameters are n_genes=number of genes, shift_result_folder, steady_state_correlation_folder, cell_state_probability_folder are the folders where the probablity shift from departure to destination states, correlation between steady state probability distribution of full and sampled networks, attractor probabilities will be stored.

5. Run Plots.ipynb notebook (R kernel). ggpubr, ggplot2, gridExtra,ggpubr are the required libraries. figure_folder is the parameter for the folder where the figures will be saved.

Please find the relevant files in SyntheticDataAnalysis/SampledNetworkConsistency folder.

1. Run Consistency_Analysis_PrepareAtts.ipynb notebook (R kernel). GA library is required. parent_sampled_folder= the folder where the sampled attractors will ve saved, num_sub_bn_sets= number of times the BN construction algorith will run, num_sel_bits= number of bits that will be subsampled for BN construction.

2. Run cons_sampled_network_constructBNs.sh from command line. The configurable parameters are Attractors_folder =the folder where the BN construction will read attractors, output_folder= where the results from BN construction algorithm will be saved, n_sets= the number of runs for the same attractor set for consistency analysis, num_bns=Number of times the BN construction algorithm will run.

3. Run PrepareDataForPlot.jl from command line. The required packages are FileIO, JLD2, DelimitedFiles, DataFrames. attractor_folder= the folder where the attractors will be read from, intervention_folder= the folder where the intervention results will be read from, shift_result_folder=the folder where the probability shift results will be written.

4. Run Plots.ipynb notebook (R kernel). ggplot2, ggpubr are the required libraries. Configurable parameters are figure_folder=the folder where the plots will be saved, prob_shift_folder=the folder where probability shift values will be read from.

Please find the relevant files in ProB_Monocyte_Example folder.

1. Download data from here.

2. Run data_imputation.ipynb (R kernel) notebook. Required libraries are scRecover and BiocParallel. Its output is the imputed scRNAseq data.

3. Run prepareAttractors.ipynb (R kernel) to prepare the input for the BN construction algorithm. The required libraries are ggplot2, dplyr, GA. The configurable parameters are departure_cell=the source cell type that is aimed to transdifferentiate from, destination_cell=the cell type that is targeted. n_cell_pairs=the number of pairs (one cell from desired cell type and one cell from undesired cell type), parent_sampled_folder= the folders where the attractors will be saved, num_sub_bn_sets number of subsamples in building sampled network approach, num_sel_bits= number of genes that will be in building the sampled networks num_bits=total number of genes.

4. Run run_bn_construction_prob_mono.sh from command line to construct BNs. Configurable parameters are Attractors_folder=the folder where the attractors will be read from, output_folder=the folder where the built BNs will be stored, n_sets=the number of cell pairs that the BNs will be built on. num_bns=the number of times that BN construction algorithm will run.

5. Run saveJuliaParameters.sh from command line to save Boolean Network rules. Configurable parameters are attractors_folder=the folder where the attractors are read from by BN construction algorithm, BNfolder=the folder where the script will find built BNs, the parameters_folder = the output folder where the BN rules are written.

6. Run getAttractors.sh from command line to save the steady state distribution. Configurable parameters attractors_folder=the folder where the steady state distributinos will be written into. parameters_folder=the folder where the BN rules will be read from.

7. Run getInterventions.sh from command line to save the steady state distribution after interventions. Configurable parameters intervention_folder=the folder where the steady state distributinos will be written into. parameters_folder=the folder where the BN rules will be read from.

8. Run Save_Probability_Shifts.R to compute probability shifts after intervention. Required libraries are stringr and GA. Configurable parameters are intervention_folder= the folder where the steady state distribution after intervention will be read from, nbits=the number of nodes in the Probabilistic Boolean Network (PBN), nsamples=number of PBNs.

9. Run PlotShift.R to visualize probability shift after interventions. Required libraries are TripleR and RobustRankAggreg and dplyr. Configurable parameters are n_sets=the number of cell pairs that the analysis has been performed on.