TissueNet is a research project started in April 2023 in collaboration with Prof. Dr. Dominic Grün, Prof. Dr. Ingo Scholtes and Dr. Leo Rasche. Single-cell RNA sequencing (scRNA-seq) has made remarkable progress during the last decade and allows now to measure a huge amount of transcriptomes of cells. Using this, researchers can characterize cell types, investigate differences in tissue composition and gene expression. In order to measure additional molecular layers some other methods have been developed, e.g., chromatin accessibility (scATAC-seq) or cell-surface proteins(CITE-seq). But commonly distinct molecular layers are measured individually and there are only a limited number of multimodal techniques which can be used more routinely. Another task is to integrate datasets across patient cohorts to investigate the development of diseases.
Integrating such multimodal data is still a challenging task and currently obtains only limited resolution. Hence, the aims of this project are to improve multimodal integration of datasets, modelling intra- and inter-cellular network dynamics, and investigate the effects of one of the most innovative anti-cancer strategies on a cellular level.
Our contributions to this project are:
- Representation learning: Achieve a joint representation of different measurement methods and increase their quality to understand disease-related perturbations of cellular states and capture the biological variability across all cell types
- Multimodal deep learning models: Preservation of as much information as possible from multimodal data
- Deep metric learning: Facilitate cell state annotation of spatial imaging data at high resolution
- Semi-unsupervised learning: Fusion of different modalities and spot yet hidden cell types
Project leaders: Prof. Dr. Dominic Grün, Prof. Dr. Ingo Scholtes, Prof. Dr. Andreas Hotho, Dr. Leo Rasche
The following persons are involved in this project: