Presenter: Aidan Daly, Research Fellow, Systems Biology
Topic: Image registration techniques for processing spatial multi-omics data
Abstract: Recently developed techniques for generating spatially-resolved measurements of the transcriptome and proteome have the promise to provide a deeper insight into the pathology of progressive diseases such as amyotrophic lateral sclerosis (ALS). Due to the experimental demands of these protocols, however, it is often impossible or infeasible to gather complete multi-omic data from a single tissue sample, and even if it were, such single-sample analyses would give limited epidemiological insights. In this talk, I will review some of the challenges in registration — the transformation of two or more sets of data into a common coordinate framework — of imaging-based omics data that are required to integrate data from different imaging modalities and individuals prior to downstream analysis. I will discuss how the nature of microscopy images require differing approaches from natural image registration, and how these differences have informed the design of a novel approach for learning over whole-slide images (WSIs).