might take naive approaches or non-clinically relevant approaches,” he said.
Guinney said this challenge, with its hundreds of thousands of samples, is one area in which the idea of big data and biomedical research don’t always have “real meaning” to each other.
Sometimes there just isn’t enough comparable biomedical information (such as a particular type of tumor or cell) available to make big data analysis possible.
“One could talk about big data in terms of these numbers of points we can measure that are in the billions, but ultimately we’re constrained by the number of samples we have,” Guinney said.
Other DREAM (Dialogue for Reverse Engineering Assessments and Methods) Challenges don’t have that same back storage of data to examine in the same way. Also, not all DREAM Challenges focus on cancer or imaging. One current challenge, called the Respiratory Viral DREAM Challenge, is trying to identify markers that protect some people against viral respiratory infections and not others.
The first challenge opened in 2006. The organization is now partnered with Sage Bionetworks and includes various other partners for other challenges. For example, the other Digital Mammography Challenge partner organizations include the National Cancer Institute, the FDA, Group Health Research Institute and others.
The challenges can prove successful because they harness rivalries to solve problems, said Guinney.
“[The goal is] having a very high impact result for the field, rather than just having a prize and people walking away and that being the end of it,” he said.