Quantification of TDP-43 inclusions in images of neurons
Advisor(s)
Dr. Jane Wu; Northwestern University
Discipline
Computer Science
Start Date
21-4-2021 9:30 AM
End Date
21-4-2021 9:45 AM
Abstract
TAR DNA-binding protein 43 (TDP-43) accumulates in the neurons and glia and has been linked to neurodegenerative diseases like frontotemporal dementia, amyotrophic lateral sclerosis (ALS), and Alzheimer’s disease. The project aims to quantify immunostaining signals of TDP inclusions for the nuclear position, type of staining patterns (homogeneous or granular/punctate), and subcellular distribution patterns in neurons. A library of images of neurons was compiled from neuronal morphology textbooks with immuno-histochemical or hematoxylin and eosin (H&E) staining at different magnifications. These images will be used to roughly estimate the precision of the algorithm. The algorithm is being developed with Image J, a Java-based image processing program. In the future, a machine learning model can be trained to recognize immunostaining signal types and distribution patterns. This will allow quantitative analysis of neuronal images and high-throughput quantification of TDP inclusions for future studies involving this protein.
Quantification of TDP-43 inclusions in images of neurons
TAR DNA-binding protein 43 (TDP-43) accumulates in the neurons and glia and has been linked to neurodegenerative diseases like frontotemporal dementia, amyotrophic lateral sclerosis (ALS), and Alzheimer’s disease. The project aims to quantify immunostaining signals of TDP inclusions for the nuclear position, type of staining patterns (homogeneous or granular/punctate), and subcellular distribution patterns in neurons. A library of images of neurons was compiled from neuronal morphology textbooks with immuno-histochemical or hematoxylin and eosin (H&E) staining at different magnifications. These images will be used to roughly estimate the precision of the algorithm. The algorithm is being developed with Image J, a Java-based image processing program. In the future, a machine learning model can be trained to recognize immunostaining signal types and distribution patterns. This will allow quantitative analysis of neuronal images and high-throughput quantification of TDP inclusions for future studies involving this protein.