Nature Communications: Benchmarking multi-omics data integration methods

The Network and Systems Biology group, in collaboration with the Marseille Mathematics Institute and the ENS Institute of Biology in Paris, publishes a benchmark comparing multi-omics integration methods in the context of cancer studies. This study compares 9 state-of-the-art joint dimensionality reduction approaches, based on matrix factorisation, for the integration of multi-omics data. The performance of the different methods are assessed in different scenarios relevant for cancer research,  from clustering simulations, associations to survival, clinical and biological annotations, to single-cell clustering.