Implementing imaging-based spatial transcriptomics at EPFL
- Abstract number
- 110
- Presentation Form
- Poster
- DOI
- 10.22443/rms.elmi2024.110
- Corresponding Email
- [email protected]
- Session
- Poster Session
- Authors
- Anjalie Schlaeppi (1, 2), Arne Seitz (1), Jessica Sordet-Dessimoz (2)
- Affiliations
-
1. EPFL - BioImaging and Optics Platform
2. EPFL - Histology Core Facility
- Keywords
hybiss transcriptomics spatial core facility
- Abstract text
Various methods exist to investigate the spatial context of specific genes to understand differences in expression across tissues and between cells. Imaging-based spatially resolved transcriptomics has become a key technology for the research community. Due to their high cost and complexity, these technologies require a large field of expertise, from molecular biology to high throughput imaging and advanced image analysis, leading to only a handful of specialized laboratories having access to these techniques. Our goal as a core facility is to democratize access to this powerful tool. Here we present our implemented workflow for hybridization-based in situ sequencing (HybISS)1, from probe design to tissue preparation, single-cell mRNA identification, and bioinformatics analysis, making it readily available for researchers.
To lower the entry barrier for this procedure, we automate the sample preparation as much as possible by using a microfluidic chamber (Lunaphore) and placing it under a widefield microscope. The multi-cycle imaging needed during the protocol can therefore be done with minimal user interaction, ensuring high-quality results even for less experienced researchers. For image analysis, we developed a pipeline to align tiles from different imaging cycles and stitch them using pre-existing software. We then use modern computer vision techniques to improve the quality and efficiency of image analyses, such as deep learning-based detection of individual RNA spots and decoding2.
We can tailor custom pipelines for our wide range of users working with mouse and human fresh-frozen tissues as well as species routinely used in fundamental research, such as zebrafish, rat, and drosophila. Here we present results obtained on zebrafish embryos.
Together, our work aims to facilitate access to tailored state-of-the-art spatial transcriptomics.
- References
1. Gyllborg, D. et al. Nucleic Acids Res. 48, e112 (2020)
2. Dominguez Mantes A. et al. BioRXiv (2024)