BoolODE: Simulated Single-Cell Gene Expression Time Series

BoolODE generates simulated single-cell expression datasets based on gene-gene interaction networks. 117 simulations were created using random ten-gene networks, each producing expression levels for 2000 cells across 800 time points. This dataset was created by Rossin Erbe to benchmark predictions of future expression states.

Use Cases

• Benchmarking future expression prediction algorithms based on simulated time-series data.
• Analyzing the impact of different gene-gene network structures (positive, negative, or no relationship) on expression dynamics.
• Testing time-series forecasting models on a controlled, simulated biological dataset.

Strengths

• 117 distinct simulations provide multiple scenarios for testing.
• Each simulation contains expression levels for 2000 cells, offering a substantial sample size.
• Data spans 800 time points per cell, enabling detailed temporal analysis.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data is simulated, not experimental, which may limit real-world applicability.

Provenance

Source

Rossin Erbe

Collection Method

Generated using the BoolODE simulation tool (Pratapa et al., 2020) with minor code modifications.