Working as a Researcher under Dr. Benjamin Bartelle at Arizona State University in the Health Systems Engineering Department, I am pursuing a project where I aim to provide a basis for the etiology of disease rather than risk detection. By working with dimensionality reduction methods such as PCA, I aimed to determine whether the morphological phenotype of the adult mouse brain can be determined by spatial gene expression. Additionally, creating an optimal representation learning method incorporating dimensionality reduction of spatially resolved biological data is another goal.
Using an adult mouse brain dataset available through the Allen Brain Atlas, we created a database through which we could run PCA and sparse filtering analysis. Initial steps included building a SQL database and registering gene expression values. Next, we visualized the gene expression using Brainrender and napari software in python to attain 2D and 3D perspectives. Some of these visuals can be seen to the right. Currently, sparse filtering analysis of the data is being done and compared to other feature learrning methods.
A 3D expression map image of the Fbln1 gene
A 2D slice of the overlap between Sagital clustering data and ground truth anatomy. Blue represents the anatomy and red represents the results of Dictionary Learning and Sparse Coding Analysis.
Visit My Github To Learn More.