Segmentation of embryonic stem cells in fluorescent images

Background

Embryonic stem cells are a fascinating model system with high biomedical potential. The maintenance of pluripotency is governed by a couple of important transcription factors. To understand the underlying regulatory dynamics, we quantify nuclear protein expression over time from single-cell time-lapsemicroscopy data.

Goal

In this project, we want to improve segmentation of cellular nuclei and quantification of nuclear fluorescent signals with machine learning methods. To this end, we will integrate additional information, like cell cycle information, and biological knowledge on typical nuclear shapes.

Data

Tracked single-cell time-lapse microscopy data of mouse embryonic stem cells

Methods

• Machine learning
• segmentation

Environment

You will have the opportunity to develop your project within a very interdisciplinary team, as the project will be jointly developed at the computational modeling in biology (CMB) institute at Helmholtz Zentrum and our Chair CAMP/TUM.

Contact

Diana Mateus

Carsten Marr