Mining insights from the
largest genomics data set in oncology,
the MMRF CoMMpass Study SM
T
HE MMRF COMMPASS STUDY SM
which has followed more than 1,000 patients with
multiple myeloma over the course of at least eight years,
continues to serve as a rich and unique source of data
for myeloma researchers, as confirmed by the new and
important findings recently reported from the latest
CoMMpass data analysis. In addition, to leverage this data
to the best advantage, the MMRF teamed up with Topcoder, a
crowdsourcing marketplace with the world’s largest community of
competitive data scientists and developers. and issued The MMRF
Topcoder Challenge. This is a contest in which the Topcoder community
used genomic data from CoMMpass to help determine which genetic factors may indicate a
high likelihood of early relapse. The new findings will be validated and combined with other data
sets to create the most complete picture of high-risk disease in relation to myeloma to date.
CROWDSOURCING WITH TOPCODER
The MMRF Topcoder Challenge tasked participants with creating a model that predicts the likelihood of early
disease progression based on real genetic data from the CoMMpass Study. Topcoder participants used genomic
data generated by CoMMpass on each patient sample to identify the predictors for early disease progression. The
MMRF recently announced the results of the contest and awarded $30,000 in prizes to the top five contestants.
To advance the understanding of genetic factors that predict early disease progression, the MMRF is looking to
validate these findings using other data sets, with the goal of creating hypotheses to study in the clinic.
In addition to being examined in The MMRF Topcoder Challenge, the CoMMpass Study data were analyzed
by cross-disciplinary teams in the DREAM Challenge, another crowdsourced competition organized
by Sage Bionetworks. A preliminary analysis of the DREAM Challenge was presented at the American
Association for Cancer Research (AACR) 2017 Annual Meeting in Washington, DC, in April.
6
T H E M M R F.ORG