Informatics and hypothesis‐driven research

Open Access
Publisher Website
Google Scholar
Abstract
The amassing of enormous data sets in genomics, proteomics and imaging has led a number of scientists to envision a future in which automated data‐mining techniques, or ‘data‐driven discovery’, will eventually rival the traditional hypothesis‐driven research that has dominated biomedical science for at least the past century. It is no surprise that prominent scientists have expressed their scepticism—to say the least—about this point of view (Allen, 2001). However, I believe that framing the debate in terms of hypotheses versus informatics, with the subtext of man versus machines, misses an important point: currently available informatics techniques can greatly assist traditional hypothesis‐driven research, but only if investigators slightly alter their practice to take advantage of this opportunity. For example, informatics tools exist that can assist investigators in formulating, assessing and prioritising their hypotheses. Many hypotheses are, in fact, straightforward extrapolations from current findings: for example, knowing that apolipoprotein E4 is a risk factor for Alzheimer's disease, …