We develop computational methods to combine spatially resolved information on tissue morphology with in situ RNA sequencing and protein detection. Using Deep Learning we can distinguish e.g. normal tissue from tumor tissue, and large-scale alignment of serial whole-slide images allows us to combine information from multiple protein detection methods. We also develop tools for efficient visualization and data interaction, at multiple scales. In the long-term perspective, we believe TissUUmaps will enable better diagnostics, prognostics, and treatment.
Here we showcase examples of the use of the TissUUmaps interface displaying different kinds of data such as spage2vec for unsupervised detection of spatial gene expression signatures, and an example of prostate cancer WSI visualization showing the results of AI based grading.
We have created a series of tutorials you can visit: tissuumaps.research.it.uu.se/howto.html . Levels beginner to advanced, from using the demos to setting up your own copy to setting up a server with TissUUmaps in it. More tutorials coming continuosly