AI-based multi-color ciliary image analysis upon expansion microscopy

In the past, we have developed computational methods to denoise and restore image data. Very recently, we have developed µSplit, a method for unmixing superimposed image channels in microscopy data. Here, our objective is to (1) advance µSplit’s capabilities, improving its performance, enabling the separation of more than two superimposed structures, and expanding its tolerance for varied intensity levels. These refinements of µSplit, while also be making it fit for application on 3D imaging data of ciliated cells and tissues. The project will involve the development of open-source tools that will make our improved methods available to life science users. (2) One concrete application aim is the separation of NHS ester and Bodipy labels in Expansion Microscopy (ExM) data. This will bring EM-like image quality to ExM at lower costs and higher efficiency. The freed-up channels can then be used to image specific ciliary components or other cell components of interest.

Fellow profile: Master degree in Computer Science, Mathematics, Physics, Engineering, Bioinformatics or related field. This project is suitable for computationally trained students with a keen interest in delving deep into the field of Machine Learning and AI, but will also offer ample opportunity to collaboratively solve real-world problems in cilia biology.