classifieR: an interactive web interface for the molecular classification of colorectal cancer from RNA-sequencing data


Year:

Session type:

Theme:

Gerard Quinn1,Tamas Sessler1,Wendy Allen1,Sarah Maguire1,Phillip Dunne1,Darragh McArt1,Harper VanSteenhouse2,Peter Gallagher1,Andrea Lees1,Dan Longley1,Bruce Seligmann2,Mark Wappett1,Simon McDade1
1Queen's University Belfast, Belfast, UK,2BioSpyder Technologies, Carlsbad, US

Abstract

Background

Colorectal cancer (CRC) is the 3rd most common form of cancer worldwide with ~700,000 deaths per year. Next generation sequencing (NGS) is leading the drive towards personalised precision cancer medicine. However, problems with data analysis of Next Generation Sequencing (NGS) data present difficulties in translating research into clinical assays. Currently colorectal patient samples can be stratified into distinct molecular subgroups based on gene expression, however this requires an experienced bioinformatician, this bottleneck can additionally prevent the adoption of these classifiers in the clinic.

Method

classifieR is developed in R with Shiny. Available for both Windows and MACOS to handle larger datasets. classifieR allows users to upload their own RNA sequencing data and specifically modified R packages CMSclassifier, CRISclassifier, DoRoTheA and MCP-counter are run in the backend.

Results

We developed the classifieR app, a dynamic and interactive web interface which allows for the characterization of RNA sequencing data into molecularly defined subtyping algorithms of colorectal cancer - providing CRIS subgroup, CMS subtype, an estimation of immune populations and transcription factor activity for each sample. The app can normalize raw count data through DeSeq2 and generate CRIS, CMS, MCP-counter scores and DoRothEA transcription factor activity scores for a large number of samples in a short space of time. Data can also be visualized on each page.

Conclusion

classifieR provides a framework which enables labs without access to a dedicated bioinformation to get information on the molecular makeup of their samples, providing an insight into patient prognosis and druggability. This application is freely available online.