Downstream evaluation — measures how reproducibly perturbation profiles retrieve their matched replicates, quantifying profile quality and similarity.
Tools
The Cytomining ecosystem provides a suite of production-ready tools for every stage of the image-based profiling workflow.
For upcoming capabilities and in-progress projects, see the Roadmap, which includes our experimental tools under active development. For a focused plan across OME-Arrow, single-cell scale efficiency, and AI-ready workflows, see the focused development initiative.
Quality control step — flags and removes low-quality cells before profile processing to prevent artifacts from propagating downstream.
Interactive exploration — view and inspect morphology profiles alongside their source cell images directly in Jupyter notebooks.
Upstream ingestion — converts raw CellProfiler, DeepProfiler, and IN Carta outputs into scalable, analysis-ready Parquet tables.
Alternative feature extraction — deep learning embeddings from raw microscopy images, bypassing classical segmentation-and-measurement pipelines.
Core processing pipeline — aggregates, normalizes, and feature-selects morphology profiles for downstream analysis.