We all know the story of Hansel and Gretel, abandoned by their wicked stepmother in the woods, who leave a trail of breadcrumbs to find their way home. Unfortunately, when they turn around to follow their trail, birds have eaten the breadcrumbs and they’re captured by a witch. Mayhem ensues.
If you read articles such as this and this, inability to retrace steps and follow breadcrumbs is a big problem in science; enough that the Director and Deputy Director of the NIH wrote about this issue here.
In the first study, a group at the biotech giant Amgen spent a decade attempting to reproduce the findings of “landmark” cancer research studies. Out of 53 papers they could only confirm 6. Six out of fifty-three… for those enough who know how technically difficult “wet lab” bench research is, a batting average of 11% is a shocking statistic for scientific reproducibility for presumably landmark conclusions.
Our societal investment in biomedical research is far too valuable to let birds eat our research breadcrumbs.
So with the “Duke Data Service” (DukeDS for short)*, we’re building a service—accessible via the web—for dropping “research breadcrumbs”.
These breadcrumbs need to be associated with the data that they refer to, so the Service also provides services for Creating, Reading, Updating, and Deleting data via a web application in your browser.
This is early days, so we’re rapidly rampling up features such as tagging your data, annotations, search and visualizing provenance. We’re also actively engaging our early “Alpha” test users and considering features such as file sync and heavy-duty encryption for PHI.
Ultimately our aim is to help researchers “map” their research and share those maps with their colleagues.
Let’s make it hard for wicked step moms.