Data

This directory contains data files, mostly in HDF5 format.

Format of filenames

Each HDF5 file has the filename of the following pattern:

{model}_{noise_type}_noise[_{noise_level}]_seed_{seed}_{variant}.h5

Parameters

• model: dynamical model from which the data is generated, which can be one of the following:
  • lv: Lotka-Volterra
  • rep: repressilator
  • emt: epithelial-mesenchymal transition (EMT)
• noise_type: type of noise, which can be one of the following:
  • additive: additive noise
  • multiplicative : multiplicative noise
  • fixed: noise with fixed variance
• noise_level: noise level in decimal, e.g. 0.05. Omitted for fixed noise.
• seed: seed for random number generator from which the data is generated.
• variant: variant of the data, which can be one of the following:
  • raw: noise added at all time points
  • clean_x0: noise added at all time points except the first

EMT data

There are two csv files for the EMT data:

• emt_mean.csv: mean proportions of cells in epithelial, intermediate, and mesenchymal states at each pseudotime point.
• emt_std.csv: standard deviations of the proportions of cells in epithelial, intermediate, and mesenchymal states at each pseudotime point.

File structure

Each HDF5 file contains the following top-level attributes:

• noise_type: type of noise, which can be additive, multiplicative, or fixed.
• noise_level: noise level in decimal, e.g. 0.05. Set to fixed for fixed noise.
• rng_seed: seed for random number generator from which the data is generated.
• param_values: parameter values used to generate the data for known models (Lotka-Volterra, repressilator).
• x0: initial condition
• clean_x0: whether the initial condition is noise-free or not. 0 if noise-free, 1 otherwise.

Each HDF5 file contains the following top-level data groups:

• train: training data
• valid: validation data
• test: test data

Each of the three data groups above is organized as follows:

• Attributes:
  • t_span: time span of all time series samples in the dataset, e.g. [0, 10]
  • t_step: time step of all time series samples in the dataset, e.g. 1.0
  • num_samples: number of time series samples
• Data groups: there are num_samples data groups, each corresponding to a time series sample. Each sample is named by sample_{i}, where i is the index of the sample, and has the following datasets:
  • t: time points of shape (num_time_points, )
  • x: values of all state variables at each time point of shape (num_time_points, num_states)