Cytomining

buscar

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

buscar is a Python package for reproducible hit calling in high-content screening. Rather than averaging cells into population-level summaries, it operates on individual cell distributions to preserve biological heterogeneity and identify perturbations with on-target morphological signatures.

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 →

copairs

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

copairs is a Python package for evaluating the quality of morphological profiles by measuring how well a perturbation’s profile can be retrieved relative to controls. It implements mean Average Precision (mAP) and related metrics widely used in the image-based profiling community.

Key capabilities:

Define matched pairs based on experimental metadata
Compute mean Average Precision (mAP) for retrieval assessment
Evaluate intra- vs. inter-group morphological similarity
Scale efficiently to large screening datasets

View documentation →

coSMicQC

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

coSMicQC (Single-cell Morphology Quality Control) identifies and removes low-quality cells from image-based profiling datasets before downstream analysis. It catches common problems such as over-segmented nuclei, poorly segmented cells, and imaging artifacts.

Key capabilities:

Flag over-segmented, under-segmented, and poorly focused cells
Apply threshold-based or z-score-based QC criteria
Generate summary reports of QC outcomes
Integrate seamlessly with CytoTable and pycytominer workflows

View documentation →

Publication #

bioRxiv Preprint · 2025

Stellar quality control for single-cell image-based profiling with coSMicQC

CytoDataFrame

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

CytoDataFrame extends the familiar pandas DataFrame interface to let researchers view and analyze single-cell morphological profiles alongside their corresponding microscopy images and segmentation masks — all within a Jupyter notebook.

Key capabilities:

Display cell images inline with profile data in Jupyter
Link numerical features directly to visual representations
Overlay segmentation masks for quality inspection
Built on top of pandas for full compatibility with existing workflows

View documentation →

CytoTable

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

CytoTable harmonizes output from different high-content image analysis tools — including CellProfiler, DeepProfiler, and IN Carta — into a single, analysis-ready format. It scales to large datasets using Apache Parquet and DuckDB under the hood.

Key capabilities:

Convert CellProfiler SQLite, CSV, and other formats into Parquet
Harmonize schema differences across analysis tools
Scale to datasets with millions of single cells
Produce outputs compatible with pycytominer and AnnData workflows

View documentation →

DeepProfiler

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

DeepProfiler uses deep neural networks to extract morphological features directly from raw microscopy images, bypassing traditional segmentation-and-measurement pipelines. It is designed for high-throughput screens where deep learning representations outperform classical feature sets.

Key capabilities:

Train and apply convolutional neural networks for feature extraction
Support for EfficientNet, ResNet, and custom architectures
Crop and embed single cells from large microscopy images
Produce embeddings compatible with pycytominer and downstream profiling workflows

View on GitHub →

iceberg-bioimage

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

iceberg-bioimage is a Python package that catalogs bioimaging data using Apache Iceberg. It scans image stores across formats, publishes structured metadata tables, and exports layouts compatible with the Cytomining profiling ecosystem — bridging raw image archives and downstream analysis pipelines.

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 →

OME-arrow

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

OME-arrow brings Open Microscopy Environment (OME) specifications to Apache Arrow, enabling microscopy images to be stored directly in data tables alongside their metadata and derived features as multilayer structs. This makes bioimaging data fast to query, easy to share, and compatible with modern tensor-based ML workflows.

Key capabilities:

Store images, metadata, and derived features together in a single table
Support for TIFF, OME-Zarr, NumPy, and Parquet source formats
Lazy reading and region-of-interest (ROI) access for large datasets
Tensor-focused output compatible with PyTorch, JAX, and DLPack
Visualization integrations for matplotlib, PyVista, and Napari

View documentation →

pycytominer

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

pycytominer is the core Python package in the Cytomining ecosystem. It provides a clean, composable API for processing single-cell morphological profiles produced by tools like CellProfiler.

Key capabilities:

Aggregate single-cell data to well- or treatment-level profiles
Annotate profiles with experimental metadata
Normalize features using population-level statistics
Select high-quality features and remove noise
Output analysis-ready profiles in standard formats

View documentation →

Publication #

Nature Methods · 2025

Reproducible image-based profiling with Pycytominer

zedprofiler

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

zedprofiler is a CPU-first toolkit for extracting morphological features from 3D volumetric microscopy images. It is designed for high-content and high-throughput workflows where classical segmentation-and-measurement pipelines need to scale to single-cell features in z-stacks.

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 →