by-experiment-results

What problem does it solve?

Ingest wet-lab readouts and calibrate in-silico predictions against ground-truth lab outcomes to close the BY wet-lab loop.

Core Features & Use Cases

• Ingest lab readouts and join them to per-design in-silico features to form a canonical calibration dataset.
• Compute calibration metrics (Mann-Whitney U, BH correction, AUC) and generate a human-readable calibration report.
• Push validated/contradicted findings to the knowledge graph and produce an enriched dataset for the next campaign optimizer.

Quick Start

Run ingest_lab_results.py to normalize the lab batch, then diagnose_silico_vs_lab.py to generate calibration results, and finally update_knowledge_from_lab.py to push findings to the knowledge graph.