Master Thesis Project in Bioinformatics:

Clustering analysis of immune-infiltrate tumors


The Cancer Genome Atlas (TCGA) generated comprehensive and multi-dimensional (DNA, RNA, protein and epigenetic) omic data of different types of cancers, such as breast cancer, lung cancer, colorectal cancer. The study of each cancer type included hundreds to thousands of candidates. One of the standard analyses of these large scale data is clustering analysis, which aims to stratify patients with different molecular features and/or clinical outcomes. The results of clustering analysis is potentially valuable in the development of personalized medicine. 


In original TCGA publications, the clustering analysis has shown that, in some cancer types, such as bladder cancer and ovarian cancer, there is a subtype in these cancers highly infiltrated by immune-cells, which sometimes is called “immunoreactive subtype”. The immune-infiltration has great prognostic value of immunotherapy. However, the immunoreactive subtype in different cancer types are associated with different genomic features and clinical outcomes, which implies the heterogeneous compositions of these tumors.


The purpose of this project is to further classify the immunoreactive tumors into different subtypes and comparing the resulted subtypes with clinical and genomic features. Generally, you will apply different clustering algorithms to large scale omic data in this project. 


Prerequisite: Familiar with statistical language R

Further readings:



JN Weinstein et al. 2013. Pan-cancer: The Cancer Genome Atlas Pan-Cancer analysis project

Cancer and immune cells:

AJ Gentles et al. 2015. The prognostic landscape of genes and infiltrating immune cells across human cancers.

F. Pages et al. 2010. Immune infiltration in human tumors: a prognostic factor that should not be ignored

DM. Pardoll 2012. The blockade of immune checkpoints in cancer immunotherapy

For further information, please contact: Dr. Chen Meng