ResearchHIGEX

Computational framework for joint analysis of histopathology images and gene expression data

Overview

The researchers and clinicians are armed with a battery of techniques to investigate both the normal and pathologic behaviors of living organisms, at different scales and resolutions. Among these techniques (modalities), in the present project we are focusing on the histopathology imaging and genomics and their applications in oncology. They provide different perspectives on the biological processes that can be combined for a better understanding of the tumor biology and for a refined diagnosis. In contrast with histopathology, the gene expression modality is relatively young, with high throughput whole-genome technologies becoming available only in the last decade.

We hypothesize that histopathology slides contain more information than can usually be exploited in routine diagnostic work by pathologists and that combining digital pathology with gene expression data will help in extracting this information and greatly enhance the diagnostic and prognostic capabilities. At the same time, the combination of these two modalities could lead to a better understanding of the biology underpinning the tumor evolution and its interaction with the host organism.

Justified also by our own observations in colon cancer subtyping, the project aims at building the computational framework necessary to jointly investigate microscope pathology images, gene expression and clinical data, and to produce tools for (semi-)automatic annotation of the images with hints about molecular processes. The work is focussed on the specific case of colorectal cancer, but the computational framework developed will be generally applicable.

The results of this project will directly benefit to the pathologists and biologists, on one hand, and computational biologists/bioinformaticians, on the other hand. Naturally, the ultimate goal is that some of the developed techniques will find their place in everyday clinical practice.

Funding

The project is financed from the SoMoPro II programme. The research leading to the results presented in these pages has acquired a financial grant from the People Programme (Marie Curie action) of the Seventh Framework Programme of EU according to REA Grant Agreement No. 291782. The research is further co-financed by the South-Moravian Region.