Bionamic and Lund University group wins FDA and NCI multi-omics data challenge

Bionamic and Lund University group wins FDA and NCI multi-omics data challenge

SmiLe Incubator company Bionamic and collaborators at the Computational Biology Division at Lund University were recently announced as winners of the “NCI-CPTAC Multi-omics Enabled Sample Mislabeling Correction Challenge” organized by the Food and Drug Administration (FDA) and National Cancer Institute (NCI). This challenge is part of President Barack Obama’s Precision Medicine Initiative in 2015. 

Clinical and biological sample mislabeling contribute to irreproducible results and invalid conclusions and represent a major, and sometimes, overlooked challenge.

The goal of this challenge was for participants to develop and evaluate computational algorithms to accurately detect and correct mislabeled samples. The datasets used in the competition were derived from 160 tumors assayed by both mass spectrometry and RNA-sequencing (multi-omics). The organizers then deliberately introduced sample mislabeling (accidental swapping of patient samples) or data mislabeling (accidental swapping of patient omics data) at rates and in patterns previously observed in various data sets.

 The model developed in the Bionamic and Lund University collaboration yielded a perfect 100% sensitivity and specificity in the test data, outperforming academic and commercial research groups from USA, Europe and Asia.

According to Anders Carlsson, founder of Bionamic, the success was due to “the broad range of world-class knowledge in our team, and our unique ability to combine them. As is probably known to most people ever having worked in a cross-disciplinary setting, to be successful you need to be able to really understand each other. The fact that both parties in our collaboration have long histories of working with everything ranging from tumor biology to artificial intelligence (AI) is what I think helped us come out on top.

Professor Carsten Peterson: “This is another example where AI technologies enter the scene with its power. However, to reach the final percentage profound application knowledge is needed.”

A paper describing the methods developed and their implications is due to be published in Nature Medicine by the organizers together with the top performing teams.

Bionamic is a SmiLe incubator company located at Medicon Village in Lund, specialized in developing tools for integration, analysis and interpretation of life science data.

Computational Biology is part of the Department of Astronomy and Theoretical Physics, Lund University.

Contact: anders [at] bionamic.se

ReferencesNature Medicine 24, 1301–1302 (2018) for the backgound;
https://precision.fda.gov/challenges/5/view/results for the results

 

 

Bionamic