Multi-modal clustering reveals event-free patient subgroup in colorectal cancer survival

This repository contains the data and associated results of all experiments conducted in our work "Multi-modal clustering reveals event-free patient subgroup in colorectal cancer survival".

Colorectal cancer (CRC) benefits from a multi-omics based stratification in the context of survival. We demonstrate that with purposeful feature selection and unsupervised clustering, multi-omics data can stratify CRC patients on the rarely studied disease-specific survival better than unimodal data and the established consensus molecular subtypes of CRC, isolating an all-surviving group of patients. 

The file coadread_data_clustering.zip contains all the data files used in the clustering analysis.

The file results.zip contains pickled kmeans clustering objects and predicted cluster labels for every modality. This information in combination with the provided data can be used to retrieve the distribution of labels across modality-derived-clusters using the code provided in our Github repository, namely the function analyse_cluster_targets found in clusering_analysis.py file.

Note: Updated version has hidden files associated with MacOS compression removed.