Approach

Summary

Licensing status

Publication and contact information

Computational models

Community-developed computational model for breast cancer progression

A computational model for breast cancer progression developed through an open innovation challenge could lead to better prognostic tests for the disease and help guide the development of future computational models. An online server was used to host user-submitted computational algorithms for predicting disease progression based on a dataset containing expression, copy number and clinical data from about 2,000 breast cancer samples. The winning model predicted progression on a validation data set of 184 samples with higher accuracy than that of previously reported prognostic models for breast cancer. Next steps include studying the biology underlying predictive features of the model and running additional challenges in other disease areas using this framework.
Oncotype DX from Genomic Health Inc. and MammaPrint from Agendia B.V. are prognostic gene expression tests marketed for breast cancer (see DREAM team, page 4).

SciBX 6(18); doi:10.1038/scibx.2013.445
Published online May 9, 2013

Findings unpatented for both studies; licensing status not applicable

Margolin, A.A. et al. Sci. Transl. Med.; published online April 17, 2013;
doi:10.1126/scitranslmed.3006112
Contact: Stephen H. Friend, Sage Bionetworks, Seattle, Wash.
e-mail:

friend@sagebase.org
Contact: Adam A. Margolin, same affiliation as above
e-mail:

margolin@sagebase.org

Cheng, W.-Y. et al. Sci. Transl. Med.; published online April 17, 2013;
doi:10.1126/scitranslmed.3005974
Contact: Dimitris Anastassiou, Columbia University, New York, N.Y.
e-mail:

da8@columbia.edu