Deep Space Exploration: Machine Learning to Understand and Annotate Image-Based Transcriptomics
To develop machine learning algorithms to analyze image based spatial transcriptomics data.
Results & Resources
In collaboration with Steve Quake’s lab and the CZ Biohub, the Zou lab developed computational models to estimate spatial gene-gene interaction from FISH data, as detailed in their preprint. They also developed a deep learning method with Joakim Lundeberg to integrate histology images with spatial transcriptomics. Finally, to make it easier for biologists to interact with image deep learning methods, the Zou lab created an interactive platform for deep learning models for diverse types of image data called gradiohub.