BioASP: Life Sciences Reasoning Benchmarks¶

CASPO, MENECO, EXDESI, PRECURSOR, IGGY

📖 Description¶

A collection of ASP benchmarks from systems biology and bioinformatics, covering reasoning on biological networks, optimization of metabolic models, and experiment design. The encodings solve real-world biological reasoning and optimization tasks using Answer Set Programming.

⚙️ Technical Details¶

• Type: Decision, Optimization (some multi-shot usage via scripting)
• Format: ASP-Core-2 / clingo
• Tested With: clingo (various 5.x versions over time)

📊 Instances¶

• Generation: Real-world biological data
• Details:
  Each benchmark contains original datasets (biological networks, reactions, experiments) transformed into ASP facts. Instance sizes vary from small curated examples to medium-to-large real-world networks.

🔒 Data & Access (Confidentiality)¶

• Status: Public
• License: Repository-specific open-source licenses (see individual repositories)
• Sensitivity: No personal or sensitive data; biological network data only

🔗 Source & Literature¶

• Repositories/ZIPs:
  • Reasoning on signaling networks (CASPO): https://github.com/bioasp/caspo
  • Metabolic network completion (MENECO): https://github.com/bioasp/meneco
  • Experiment design (EXDESI): https://github.com/bioasp/exdesi
  • Minimal metabolic precursor sets (PRECURSOR): https://github.com/bioasp/precursor
  • Influence graph analysis (IGGY): https://github.com/bioasp/iggy
• Generator URL:
  Data preparation and instance generation scripts are included in the respective repositories.
• Reference:
  See publications cited in the individual repositories (systems biology / BioASP literature).

👤 Contact¶

• Name: BioASP contributors
• Email: See GitHub repository contacts / maintainers

💡 Miscellaneous¶

• Complexity:
  Domain-specific reasoning and optimization problems; includes NP-hard optimization tasks.
• Metadata:
  No unified metadata format across repositories; instance properties are partially documented in README files.
• Other Notes:
  All benchmarks follow a similar workflow:
  1. Transform biological data into ASP facts
  2. Combine with problem-specific ASP encodings
  3. Solve using one or multiple calls to clingo
  4. Post-process solver output into domain-specific results