Bioinformaticians from the “Bioinformatics and Modelling” (BIOMOD) research group at LIH’s Department of Oncology took part in the Proteogenomics challenge organised by the DREAM Challenges and ranked in the top-10 participating teams.
The DREAM Challenges are crowdsourcing competitions examining questions in systems biology and translational medicine. They constitute a non-profit, collaborative community and open science effort that involves contributors from across the research spectrum: research teams from universities and public research institutions, technology companies (e.g. IBM Research), not for profit organisations (e.g. Sage Bionetworks) as well as biotechnology and pharmaceutical companies.
The challenges can help scientists to get access to more powerful and adapted methods to solve their biological questions. In 2017, the BIOMOD research group at LIH participated in the NCI-CPTAC DREAM Proteogenomics Challenge (subchallenge 3), and ranked among the top-10 best performing participants. The challenge was about developing computational models that can predict the phosphorylation levels of human proteins in different cancer samples. Teams were ranked according to the quality of their predictions.
‘Participating in this challenge was a very enriching experience’, states Dr Tranchevent, post-doctoral researcher in the BIOMOD research group. ‘The challenge was announced in spring 2017 and the associated datasets were subsequently released. We then had a few months to get acquainted with the data and to propose our solution to the problem. Our model combines genomic, transcriptomic and proteomic data and uses a machine learning strategy to predict the phosphorylation levels of hundreds of human proteins. Ultimately, the goal of such challenge is to compare different modelling strategies as to better understand the mechanisms behind protein phosphorylation.’
Dr Francisco Azuaje, BIOMOD research group leader, is proud of the team’s achievement: ‘This was a successful collective effort that allowed us to further demonstrate our research capacities in a competitive international setting. ’
