Case studies

Case studies are heavier than the tutorials. Each one runs a realistic analysis end to end on a full-scale dataset, where the tutorials stay small enough to execute in CI on every change. Because these run at full scale, they are pre-rendered from committed outputs rather than executed in the build, and re-run on demand when the analysis changes.

• Learned embeddings against CellProfiler
• Reproducing across laboratories
• CRISPR knockouts
• Calling hits against genes that cannot respond
• Calling gene hits in a pooled screen
• Correcting plate position for hit-calling and profiling