I posted my first research paper with Abbasi Lab and a great team of collaborators: "Unsupervised pattern discovery in spatial gene expression atlas reveals mouse brain regions beyond established ontology."
Here's the link: https://www.biorxiv.org/content/10.1101/2023.03.10.531984v1.full.pdf)
In this manuscript, we show that unsupervised machine learning (ML) can rapidly create maps of the brain based purely on 3D gene expression data. We demonstrate this in the adult mouse brain that goes beyond established ontology. Existing maps are typically hand-drawn, and can take years of person-hours to complete. By contrast, this method (osNMF) can run in a few hours on a MacBook Pro and is potentially less biased. It is applicable to any tissue or organism. Tissue maps are important. They help us understand function, development, disease, and aging.
This new method is just one of the countless examples of the powerful combination of ML + new data generation methods. Together, they are increasing the pace of biological discovery. Ours was a relatively basic ML approach––and it leads to orders of magnitude improvement in generating tissue maps. This improvements are happening all over the place.