Can protein disorder help explain how the immune system recognizes cancer?
Proteins are often represented as three-dimensional structures, but this is not always the case. Many proteins, or regions within proteins, are “intrinsically disordered” and do not adopt a stable structure.
Our cells continuously break down proteins and present short fragments of them on their surface through molecules called human leukocyte antigens (HLA). This molecular presentation allows our immune system to sample what is happening inside a cell and are central to how the immune system recognizes cancer cells.
An interesting question is whether the physical state of a protein influences which of its fragments eventually become presented by HLA. Previous studies suggest that protein disorder is associated with protein degradation and HLA antigen presentation. However, it is much less clear whether this relationship is different in cancer cells.
The main question in this MSc project will therefore be:
Does cancer change the rules that determine which structural regions of the human proteome are sampled and presented to the immune system?
To investigate this, the student will use datasets containing experimentally measured HLA-presented peptides from cancer and normal tissues. These peptides can be mapped back to their source proteins and to their position within the protein sequence. Computational methods for predicting protein disorder can then be used to determine whether peptides originate from structured regions, intrinsically disordered regions, or boundaries between them.
A simple comparison of cancer and normal peptides is not sufficient. HLA genotype, protein expression, peptide sequence, tissue type and experimental differences can all influence which peptides are detected. An important part of the project will therefore be to develop a computational framework that takes these factors into account and allows us to test whether protein disorder contributes additional information.
We will first use large public cancer and normal-tissue immunopeptidomics datasets. Observations from these data can then be tested in smaller datasets where HLA-presented peptides have been measured in tumor and adjacent normal tissue from the same patients.
There is also the possibility to extend the project to cancer genomics. Cancer cells acquire mutations and other genomic alterations that can change the proteins they produce. We can ask whether some of these alterations also change protein disorder, and whether this is associated with changes in the peptides presented to the immune system.
A better understanding of how cancer cells select which peptides to present to the immune system could ultimately help identify new targets for cancer immunotherapies or biomarkers for cancer detection and treatment.
This project is primarily computational and would suit an MSc student interested in bioinformatics, statistics, or data science, with some programming experience in Python or R and a willingness to learn about cancer immunology. No prior knowledge of cancer biology or immunology is required.