The Integration of Machine Learning Tools into Spatial Transcriptomics Image Acquisition and Analysis
- Abstract number
- 359
- Corresponding Email
- [email protected]
- Session
- Stream 2: Machine Learning for Image Analysis
- Authors
- Dr Kwasi Kwakwa (1)
- Affiliations
-
1. EMBL-EBI
- Abstract text
Spatial transcriptomics is a rapidly advancing family of techniques for resolving the spatial distribution of gene expression in large tissue sections. Specifically, the subset of In Situ Hybridization(FISH) based techniques use a combination of direct RNA labelling and fluorescence microscopy to resolve gene expression with subcellular precision. A significant challenge in the analysis of this fluorescence data is the accurate segmentation of cells in tissue sections, where different tissue types can be constructed of cells with varying morphology as well as differing levels of spatial crowding. As part of the process of creating automated image acquisition and analysis pipelines for high throughput processing of labelled tissues, we have been investigating the use of machine learning tools for cell segmentation. I will be talking about the lessons we have learned in the process and all the challenges we see going forward.