Cell Painting is a specialized high-content imaging technique used to extract content rich phenotypic information from cellular samples. Its advantage is the capture of un-biased high-dimensional morphological data rather than focusing on specific biomarkers. The Cell Painting technique is also capable of generating large scale data suitable for more complex data interpretation tasks like multivariate data analysis in machine learning and AI-driven analysis.
Basic Concept of Cell Painting: iPSC derived podocytes with staining of important cell compartments like nuclei (blue), lysosomes (red), actin (green) and mitochondria (yellow) prepared for Cell Painting
Cluster Analysis Enriched with Reference Data: Cell Painting utilizes high number of features extracted from image processing tasks. On these phenotypic fingerprints, high dimensional data analysis is performed such as feature weighting, nearest neighbor or cluster analysis. Enriched by treatment annotations this high dimensional dataset can be interpreted and be taken for decision making.
Applications of Cell Painting in Multiple Drug Discovery Phases
The fully established yet flexible Cell Painting workflows at Evotec add value in various phases of drug discovery:
Phenotype characterization of cellular responses
In well-established workflows Evotec generates phenotypic data with diverse cell models, flexible sets of stainings, best-in-class instrumentation, and powerful image processing algorithms followed by state-of-the-art multivariate data analysis pipelines. Cell Painting has already been successfully applied in many projects e.g. in iPSC donor selection and disease progression characterization.
Cell Painting Workflow at Evotec
Cell Painting utilizes high number of features extracted from image processing tasks. On these phenotypic fingerprints,nhigh dimensional data analysis is performed such as feature weighting, nearest neighbor or cluster analysis. Enriched by treatment annotations this high dimensional dataset can be interpreted and be taken for decision making.
Target identification, relation and validation
In target validation and deconvolution assays, Evotec utilizes content rich high dimensional phenotypic fingerprint information to translate pre-existing knowledge on compounds (annotations) or genes (CRISPR constructs) into target relation. By comparison with known landmark compounds, the prediction of e.g. Mechanism-of-Action of unknown compounds or reversion of disease specific phenotypes is possible
Cluster Analysis to Discover Target Relation
Nearest neighbor analysis in high dimension allows to identify similar phenotypes and ranking by distance. By including annotated reference compounds, the mechanism of action of a close by candidate compound can be predicted. The illustration shows a two dimensional representation using t-SNE as dimensionality reduction method.
Zoomed in illustration: Known annotated reference compounds (red) are in close proximity to a candidate compound (light blue) with various replicates and concentrations.
Hit identification and compound library profiling
Cell Painting is one flavour of phenotypic screening used for hit identification including additional possibilities for hit triaging and early clustering analysis. Typically the data are used to generate large scale phenotypic fingerprint profiles for AI/ML-based compound characterization and prediction of compound activity of complete libraries (“Hit Prediction”). Every data driven or iterative screening approach can benefit from adding cell-based content rich phenotypic fingerprints using Cell Painting.
Hit Prediction Workflow at Evotec
Information of several quality-controlled datasets from the in-house fingerprint database is used for AI/ML to predict the outcome on incomplete data sets. This approach improves timelines by screening smaller decks and helps to save reagent and other material costs.
Early toxicity and safety assessments (Safety/Tox)
The early assessment of Safety/Tox related parameters with standardized and scalable, equipment allows fast turnaround times, highly efficient data generation and analysis with sensitive readouts. Thousands of different features can be assessed based on intensity, texture and granularity of each dye. Phenotypic changes can be compared with a database of compounds with known toxic effects to understand safety and potential toxicity. This makes Cell Painting a valuable approach for safety assessment.
What is Cell Painting?
Cell Painting involves staining multiple cellular compartments with fluorescent dyes, imaging them using high-content microscopes, and analyzing the resulting images with machine learning and AI techniques. It enables deep cellular profiling for drug discovery and disease research.
The main steps involved in Cell Painting include:
- Key organelles or cellular components are stained
- Phenotypic fingerprints are extracted for high-dimensional analysis
- AI-driven clustering and pattern recognition enhance data insights
What are the advantages of Cell Painting?
- Unbiased, high-dimensional data acquisition.
- Scalable and cost-efficient for high-throughput screening.
- Compatible with AI and machine learning techniques.
- Provides comprehensive insights into cellular behavior.
Why choose Evotec as your CRO for Cell Painting?
At Evotec we offer the complete Cell Painting workflow from the wet lab to complex cutting edge data analysis under one roof. We have proven track record of running successful Cell Painting campaigns in a broad variety of cell lines. A dedicated team of specialists provides state of the art image analysis, downstream data processing and data interpretation support that is tailored to answer the customers’ scientific questions.
Learn More in Our Webinar
