Roadmap on Cytomining

buscar

Mon, 01 Jan 0001 00:00:00 +0000

buscar scores perturbations directly on single-cell distributions using Earth Mover’s Distance, preserving heterogeneity throughout hit calling.

Problem: Population-level hit calling averages away biologically meaningful cell-to-cell variation, making heterogeneous responses and rare subpopulations invisible to standard metrics.

Key capabilities:

Define on-target and off-target morphology signatures from reference profiles
Score perturbation efficacy via Earth Mover’s Distance
Assess specificity with off-target scoring to reduce false positives
Preserve single-cell heterogeneity throughout hit calling
Integrates directly with Pycytominer, coSMicQC, and CytoTable workflows

View on GitHub →

iceberg-bioimage

Mon, 01 Jan 0001 00:00:00 +0000

iceberg-bioimage scans any image store into a versioned Apache Iceberg catalog that directly exports Cytomining-compatible Parquet warehouses.

Problem: Raw bioimaging archives have no standard catalog — finding, versioning, and joining images to downstream data requires bespoke scripts per lab.

Key capabilities:

Scan image stores into canonical ScanResult objects
Publish image metadata with PyIceberg for versioned, queryable catalogs
Export Cytomining-compatible Parquet warehouses for profiling workflows
Validate profile tables against microscopy join contracts
Supports Zarr, OME-TIFF, and Parquet source formats

View documentation → · View on GitHub →

OME-arrow

Mon, 01 Jan 0001 00:00:00 +0000

OME-arrow embeds images as first-class columns in Apache Arrow tables, so features, metadata, and pixel data travel together and can be queried or exported as tensors.

Problem: Images and feature tables live in separate systems — linking a numeric outlier back to its source cell requires error-prone manual joins across formats.

ZEDprofiler

Mon, 01 Jan 0001 00:00:00 +0000

ZEDprofiler extracts morphological features directly from 3D volumetric images, including anisotropic voxel spacing correction — no GPU required.

Problem: Classical profiling tools extract only 2D features, leaving organoid, cleared-tissue, and z-stack experiments without a CPU-efficient extractor.

Key capabilities:

Extract features from 3D volumetric (z-stack) single-cell images
Multi-channel fluorescence microscopy support
Anisotropic voxel spacing correction for accurate 3D measurements
Modular, extensible architecture for custom feature sets
CPU-optimized for high-throughput processing without GPU dependency

View on GitHub →