Machine Learning and Image Analysis Core

PROJECT SUMMARY/ ABSTRACT MACHINE LEARNING AND IMAGE ANALYSIS CORE The Machine Learning and Image Analysis Core of The Jackson Laboratory Nathan Shock Center (JAX NSC) uses cutting-edge data science approaches to develop novel aging biomarkers to support JAX NSC studies and the broader aging community. Our studies leverage the diversity of our mouse models to ensure the generalizability of our biomarkers and undertake outreach activities to support their adoption. The Specific Aims of the Machine Learning and Image Analysis Core are: Aim 1. Adapt a quantitative image analysis pipeline for the analysis of aged heart, liver, lung, spleen, and ovarian tissues by training and validating classifiers. This Aim continues our work from the current funding period, where we developed first-in-class digital pathology models to predict tissue aging and extensively developed new neural network architectures to support model interpretation. We now have a mature, cutting-edge framework to extend these efforts to multiple tissues, including ovarian tissue, that will be generated during the next funding cycle. Aim 2. Adapt a quantitative machine learning pipeline for the analysis of hematology signatures across ages by training and establishing classifiers. We will extend our biomarker development efforts to clinically translatable hematology analyzer data (i.e., scattergrams) from our diverse aging mouse cohorts, using machine learning approaches to develop blood biomarkers of aging. Aim 3. Develop and distribute open-source, user-friendly packages for both the quantitative and discovery pipelines with online training to the geroscience community. Once our pipelines have been trained and validated, we will make them publicly available as open source to the community. This resource sharing will stimulate their use and will allow other groups to extend and improve upon our work. The core supports aging research by generating novel biomarker resources from our studies that are designed to support a broad range of aging research activities.