sbv IMPROVER Diagnostic Signature Challenge


Authored by  J Hoeng, G Stolovitzky, M Peitsch

Published in Systems Biomedicine     

Abstract

The task of predicting disease phenotype from gene expression data has been addressed hundreds if not thousands of times in the recent literature. This expanding body of work is not only an indication that the problem is of great importance and general interest, but it also reveals that neither the experimental nor the computational limitations of translating data to disease information have been satisfactorily understood. To contribute to the advancement of the field, promote collaborative thinking and enable a fair and unbiased comparison of methods, IMPROVER revisited the problem of gene-expression to phenotype prediction using a collaborative-competition paradigm. This special issue of Systems Biomedicine reports the results of the sbv IMPROVER Diagnostic Signature Challenge designed to identify best analytic approaches to predict phenotype from gene expression data.

Our Findings

We are committed to sharing our results as the development and scientific assessment of our products advance. Read our most important findings with supporting peer-reviewed articles, presentations, and registered clinical trials.

What we have learned so far...