code and results for a TCBB paper submission

our implementation is mainly based on following packages:

python 3.7
pip install keras==2.3.1
pip install gpuinfo
pip install tensorflow-gpu==1.15
pip install gpflow==1.5

Besides, some basic packages like numpy are also needed.

• main.py: Main program for static GRN inference.
• main_ts.py: Main program for dynamic GRN inference.
• main_monocle_ts.py: Main program for realistic HSMM dataset.
• main_monocle_ts.py, main_monocle_ts.py, main_monocle_ts.py, main_monocle_ts.py: Scripts for running experiments. Paths need to be reconfigured before execution.
• synnet_and.py, synnet_andnot.py, run_synnet_and.sh, run_synnet_andnot.sh: Scripts for running the synthetic experiments.
• run_hESC200.sh, run_hHEP200.sh: Scripts for running the single-cell RNA experiments.
• draw_fig.ipynb: Notebook for visualizing static GRN inference results.
• demo_toy.ipynb: Notebook for dynamic GRN inference results on the toy dataset.
• dynamic_synthetic_analysis.ipynb: Notebook for visualizing dynamic GRN inference results on the synthetic dataset.
• dynamic_HSMM_analysis.ipynb: Notebook for visualizing dynamic GRN inference results on the realistic HSMM dataset.
• ./eval/*: Scripts for evaluation, using GeneNetWeaver and BEELINE.
• ./results/synthetic/*: Inferred network for the synthetic dataset in each time step.
• ./results/HSMM/*: Inferred network for the realistic HSMM dataset in each time step.

• infile: Path of the input file.
• outfile: Path of the output file.
• sizen: Number of training instances.
• sizem: Number of inducing points.
• gene: Number of genes.
• iter: Steps of iterations.
• ktype: Kernel type, Poly1 (linear kernel) or Poly2 (degree 2 polynomial kernel).
• lr: Learning rate.

• ./data/static/*: 5 networks and multifactorial data for static GRN experiments.
• ./data/dynamic/synthetic/*: Multifactorial and time series data of the synthetic dataset.
• ./data/dynamic/HSMM/*: Time series data of the HSMM dataset.