Technology

Morphological profiling by Cell Painting

Cells contain a plethora of information that is hidden to the naked eye. With a combination of organelle-specific fluorescent dyes and high-content imaging, Cell Painting enables capturing a high magnitude level of detail that translates into unique cellular signatures.

Screening

We use Cell Painting when screening disease models for repurposed drugs and discovery of novel leads. The generated rich data and state-of-the-art AI methods enable us to perform iterative, intelligent screening.

Predictions

Genetic, chemical or biological perturbations induce distinct cell morphological changes. We exploit this to extract individual profiles of each of the perturbations, and use them to predict mechanism of action, target or cellular toxicity.

Artificial Intelligence

Phenaros is an AI-first company and a key focus is the capturing, storing, and processing of data to enable the use of sophisticated analytics methods, such as various types of artificial intelligence. We have over time collected a large and valuable knowledge base that enables us to compare and relate new experiments to e.g. effects of well-studied chemical and genetic probes.

Designing Experiments

Choosing the number of replicates, concentrations and constructing robust experimental layouts has a large effect on the resulting data. We design our experiments with AI to reduce the risk of errors and maximize the quality of the obtained data.

Data Insights

We analyze data using unsupervised and supervised methods, both extracted morphological profiles and the microscopy images. We routinely use methods such as Deep Neural Networks to maximize the value of the analyzed data and make robust and accurate predictions.

Automation

We believe in data robustness, data reproducibility and throughput. That is why we have built a robotised laboratory specifically tailored to perform morphological profiling experiments by Cell Painting.
‍
Our automation team has engineered and programmed our robots for precise scheduling of experiments, ensuring the highest reproducibility between experiments.

Image Analysis

We generate thousands of images per experiment, which translate into millions of cells analysed, and billions of morphological features extracted. To support such high amounts of data, we have built a powerful information technology infrastructure for automated image and data storage, image analysis, data analysis and data visualisation.

Highest Quality

This unique combination of hardware and software automation allows us to process thousands of samples in a seamless, ordered and robust manner, ensuring the highest experimental quality.

Data-rich screening driven by AI

Screening with Cell Painting enables studying the morphological changes when cells are exposed to compound libraries in a new dimension than previously possible.
‍

Defining the effect space

By establishing a desirable phenotype (e.g. "healthy") and a non-desirable ("disease") it is possible to define the morphological effect space for a screen. This could for example be two different cell states with distinct morphological features.

Reversing the phenotype

The objective of the screen is then to identify compounds that reverse the undesirable phenotype and drive cells towards the desirable phenotype. With the data-rich content from Cell Painting technology, this methodology can be used together with AI to carry out efficient, iterative discovery cycles.

How does it work?