Related Experiment Video
Updated: May 8, 2025
10:52
A High-content Imaging Workflow to Study Grb2 Signaling Complexes by Expression Cloning
Published on: October 30, 2012
10.6K
A highly efficient, scalable pipeline for fixed feature extraction from large-scale high-content imaging screens
Gabriel Comolet1, Neeloy Bose1, Jeff Winchell1
1The New York Stem Cell Foundation Research Institute, New York, NY 10019, USA.
Iscience
|December 25, 2024
Summary
Artificial intelligence (AI) applied to cell imaging aids disease and drug discovery. Our new ScaleFEx pipeline efficiently extracts key features from large image datasets for AI analysis.
Area of Science:
- Computational biology
- Bioinformatics
- Cellular imaging
Background:
- Artificial intelligence (AI) in morphological profiling of cells from high-content imaging (HCI) screens shows promise for identifying disease states and drug responses.
- Large-scale HCI screens are crucial for detecting subtle differences between cell populations and accounting for experimental variations.
- Analyzing high-dimensional HCI datasets efficiently while retaining interpretable and predictive features presents a significant challenge.
Purpose of the Study:
- To introduce ScaleFEx, a novel, memory-efficient, open-source Python pipeline for feature extraction from HCI datasets.
- To enable efficient analysis of large-scale HCI data using minimal computational resources or scalable cloud infrastructure.
- To facilitate the use of AI in identifying disease phenotypes and drug responses from cellular imaging data.
Main Methods:
- Development of ScaleFEx, a Python pipeline for memory-efficient feature extraction from high-content imaging datasets.
- Integration of ScaleFEx with AI models for analyzing phenotypic shifts in drug-treated cells.
- Application of ScaleFEx to public HCI datasets for validation and demonstration.
Main Results:
- ScaleFEx successfully extracts biologically meaningful features from HCI datasets with minimal computational resources.
- The pipeline enables AI models to identify phenotypic shifts in drug-treated cells and rank interpretable features.
- ScaleFEx is applicable to public datasets, demonstrating its broad utility.
Conclusions:
- ScaleFEx is a powerful tool for accelerating the discovery of disease-associated phenotypes and novel therapeutics through efficient AI-driven image analysis.
- The open-source pipeline democratizes access to advanced computational tools for cell imaging research.
- ScaleFEx enhances the predictive power and interpretability of AI models in morphological profiling.
