Deep Phenotyping Advanced Technology Facility

Introduction

Professor John Le Quesne Head of Deep Phenotyping Facility john.lequesne@glasgow.ac.uk

Our team design assays to answer scientific questions via the generation of data-rich images and bespoke image analysis methods to measure gene expression at multiple levels in intact tissues. We combine experience in the use of antibodies and RNA detection technologies to develop assays which probe biological areas of interest. This is achieved by working closely with our collaborators, which is essential to obtain the desired results from these challenging methods. 

We routinely develop in situ assays to answer questions about the spatial context of the tumour microenvironment, including the immune system, plasticity of epithelial cells and fibroblasts, and cell-cell interactions. 

Key technologies 

Multiplex Immunofluorescence 

Using our Ventana Discovery Ultra autostainers in combination with the Akoya PhenoImager HT, we routinely detect up to 6 genes of interest on multiple tissue sections including tissue microarrays (TMAs), allowing us to answer complex questions about tumour biology in a high-throughput manner. We are leaders in this technology, and have collaborations with leading industry providers as well as numerous academic partners. We believe multiplex assays will form the basis of the next generation of clinical biomarkers, and are working towards translating our assays into a clinical setting. 

Over 2024, we have worked with 13 research groups with this technology, and we have several manuscripts in preparation. We have developed panels for human prostate cancer (Leung, Campbell), human pancreatic ductal adenocarcinoma (Chang), human lung adenocarcinoma (Le Quesne, Bushell, MacVicar), human colorectal cancer (Edwards, Roxburgh), human breast cancer (McPherson), mouse models of hepatocellular carcinoma (Bird), cholangiocarcinoma (Braconi), and both human and mouse models of malignant mesothelioma (Chalmers, Murphy). 

High Plex Immunofluorescence 

Using our Akoya PhenoCycler Fusion, we can detect up to 100 proteins of interest within tissue spatial context. Our expertise with this platform has led to several large in-depth phenotyping collaborations, including our involvement in  the SAMBAI cancer grand challenge. This project is a multi-centre international effort to tackle cancer inequities in minority and socially deprived patient populations. In addition, we have collaborations with the Pearson lab at Cardiff University, focussing on fibroblast subtypes in human prostate cancer, and locally with the Chang, Inman and Le Quesne labs focussing on fibroblast subtyping and immune phenotypes in human malignancies.  

We have a methods paper detailing our in-depth validation protocol in process, and our custom PDAC panel has been presented at several meetings. This panel of 43 markers simultaneously detects expression of numerous markers of epithelial and fibroblast plasticity, new therapeutic targets, and highly detailed immunophenotypes. A murine version of this panel is in preparation. 

Xenium 

Our 10x Genomics Xenium platform enables detection of up to 5100 genes of interest in a single tissue section. The custom gene expression chemistry has been used by our collaborators to develop panels detecting multiple species simultaneously in parasitically infected tissues, and also to detect single nucleotide variants. Furthermore, the Xenium can be combined with PhenoCycler Fusion, enabling multi-omic detection of RNA and protein within the same tissue section. Of note, the Xenium is part of our SAMBAI cancer grand challenge work package, and data from our first mouse custom gene expression panel has been submitted to Nature by the Sansom lab. 

Visiopharm 

Visiopharm is our image analysis platform of choice. We have developed bespoke deep learning pipelines for several diseases enabling high throughput analysis of the images generated by our spatial platforms. We offer training for collaborators wanting to generate their own data, or we design custom workflows and run them as part of our deep phenotyping service. We are in the process of writing a manuscript for a novel deep learning classifier we have developed using the deep learning tools within Visiopharm, which accurately predicts cellular phenotypes, without the need for accurate cell boundary classification. 

John Le Quesne also leads the Deep Phenotyping of Solid Tumours research group.

Email