Simplifying transcriptomic profiling and molecular subtyping of colorectal cancer

Abstract

We hypothesised that the use of a high-throughput RNA-sequencing (RNA-seq) technology paired with an easy-to-use data analysis application could open the field of transcriptomic subtyping to a broader user base. TempO-Seq is particularly amenable to clinically relevant samples such as FFPE and remains an extremely high throughput sequencing technology. This thesis investigates the utility of the TempO-Seq sequencing platform compared to other sequencing technologies on its ability to sequence relevant patient data and samples assigning samples to colorectal cancer transcriptomic subtypes. In addition, we find the potential need to develop a platform-agnostic classification system that can classify patients across multiple sequencing platforms.

To simplify and expedite this data analysis workflow, we developed classifieR, available across any platform as it is hosted online. This application was designed to be as easy to use as possible, allowing non-expert users to annotate their uploaded RNA-seq data across two disease types functionally. The tool includes pre-processing of the data through DeSeq2 normalisation and commonly used subtyping algorithms for colorectal (CMS, CRIS) and breast cancers (PAM50). classifieR also annotates samples with immune and stromal cell population estimators such as MCP-counter and xCell and other analyses such as estimation of transcription factor activity (DoRothEA) and pathway analysis through single sample Gene Set Enrichment Analysis (ssGSEA). The classifieR application improves accessibility to cutting-edge transcriptomic analysis to labs and individuals without access to a dedicated bioinformatician. In addition, with the generatR dashboard being developed, we can host infrastructure providing access to easy-to-use tools that we and other groups can use in the future.

Thesis is embargoed until 31 July 2028.

Date of Award	Jul 2023
Original language	English
Awarding Institution	Queen's University Belfast
Sponsors	BioSpyder
Supervisor	Simon McDade (Supervisor), Daniel Longley (Supervisor) & Bruce Seligmann (Supervisor)