We are Microscopy Image Analysis Experts

Today’s forefront of drug discovery is empowered by high quality and structured data. TK Analytics helps you manage your imaging data flow, analyze and organize it, and make it accessible when needed.

Our Services

Image and Data Analysis

Imaging datasets are extremely rich with data. Besides routine analyses we also look for new paths in image and data analysis to get more insights into biology. Hint: we’re good friends with AI.

Self-service Data Applications

We are scientists building data applications for scientists, so you can process and visualize your data with a click of a mouse.

Screening Image Management

Your datasets, organized, searchable and enriched with public or private databases and AI predictions. That’s how we supercharge phenotypic screening in drug discovery.

Image Recognition

‎

The first step to discovery is extracting data and features from the images. We use state-of-the-art tools to do routine analyses, or develop custom algorithms to measure what hasn’t been measured yet.

CellProfiler

CellProfiler is the most versatile and powerful microscopy image analysis tool and we use it in most of our projects. The suite has a large collection of analysis modules, enabling advanced segmentation, masking, image conversions, intensity and shape measurements and texture analyses. It is also extendable, so we can also use one of the many available plugins that also allow us to work with other types of objects, such as yeast, roundworms or human tissues. We can also use AI to improve the segmentation, extract other features or classify cells or objects based on machine learning.

Cell painting

Cell painting is a powerful method for phenotypic profiling. We use the standard pipeline for image and data analysis, which we carefully enriched with quality control checks. After the features are extracted we process the data and create rich and interactive visualizations.

Data applications

When analyses become routine scientists need to be more efficient with data analysis. To empower them we develop self-service data applications which helps them explore and analyze data on their own. The applications run in a web browser, are beautifully designed and user friendly (or better said – scientist friendly).

Custom image analysis algorithms

If the analysis algorithm for your particular challenge hasn’t been invented yet, we invent it! We’ve invented many custom algorithms, to quantify different punctate distributions, endoplasmic reticulum structure, or even describe subcellular localizations. You can find some in case studies. When inventing new algorithms you’re measuring what hasn’t been measured before. It’s a special moment, for you and for us.

HCS Image Management

‎

Keeping your data organized, annotated and searchable is crucial to make the best use of all the data you acquired throughout years. Not just images, but also other experiments, such as sequencing, proteomics, and functional studies. We provide a full data management service, including data flow analyses, building a digital culture within the organization and related technical services. In the end your data will be FAIR – Findable, Accessible, Interoperable and Reusable.

Data management project overview

Each data management project starts with a careful analysis of the existing situation. We identify data flows, bottlenecks, existing infrastructure and requirements. Based on that we develop a data management strategy. This includes building a digital culture within the organization – a crucial step to make scientists feel empowered by the digital tools. That’s why this is usually a joint effort between your and our team.

OMERO - an image management platform

OMERO – Open Microscopy Environment is our go-to platform for image management. It’s highly customizable, user friendly and follows FAIR principles. It brings all your images to the fingertips of the scientists. They can quickly preview them, search by metadata, view plate layouts, or prepare report ready figures. We also develop custom tools to automate certain tasks, such as data import or finding representative images.

AI to the Help

‎

AI and deep learning made a revolution in image analysis. Microscopy is no different – AI is being used, among others, to analyze tissues, do better segmentation, classify or compare phenotypes or extract different features.

Ilastik

We routinely use ilastik for classification of phenotypes. It’s a supervised classification tool – you select a few examples of each class and then the algorithm finds similar phenotypes. It runs from CellProfiler and usually you would need to segment the cells before classification.

CellPose

Cellpose is an AI algorithm for cell segmentation. It doesn’t need nuclei to find the cells, and often it works better than traditional segmentation algorithms. The advantage is also that it works on different modalities. We use it to segment fluorescence and brightfield images.

DeepProfiler

DeepProfiler takes cell painting to the next level. Besides thousands of features extracted in the standard cell painting pipeline, this tool extracts even more features that are important to distinguish similar phenotypes. It’s a very powerful tool that we really like!