Share Alike

A survey finds researchers are less willing to share full study protocols or datasets than in the past.

Full-text access for registered users only. Existing users login here.
New to GenomeWeb? Register here quickly for free access.

Over the last 14 years,

Over the last 14 years, greater than 80% of the more than 1650 clients of Kinexus Bioinformatics Corporation have opted to have the results from proteomics analyses of their tissue and cell lysate specimens made freely available for a discounted service fee. To receive the reduced price, clients had to fully supply details about the nature of the samples, including treatment details. In turn, after a minimum 6 month hold, Kinexus has posted the measurements from these antibody microarray and immunoblotting analyses on their KiNET-AM and KiNET-IB open-access, on-line databases. Such a willingness on the part of researchers to share data even before publication has also permitted the growth of government-sponsored gene sequence and gene expression databases, which have proven to be invaluable resources for the biomedical research community.

Often, researchers have the time and resources to follow up on only a few of the interesting observations that they generate from their experiments. In the past, when a principal investigator retires from an academic laboratory, three decades or more of laboratory results was usually discarded, with probably only a very small percentage of the data ever being published. When such experiments were performed properly with reliable methods and reagents, this is a tremendous waste of resources. Even the negative results from studies have real value.

Ideally, published scientific papers represent distillations of large amounts of data from multiple experiments. However, unpublished, old data re-examined with new insights and from different angles can generate and even test newly formulated hypotheses. The larger the data set that can be interrogated, the more revealing such meta-analyses can be. With increasing development and adoption of higher throughput technologies, it will become even more important to create and maintain open-public repositories of experimental data. Queries of such databases will lead to better hypotheses generation for the design of future experiments and improve our understanding of complex biological systems. In the not too distant future, perhaps such databases will be even more useful to the synthetic intelligence systems that we design, like IBM's Watson for example, for biomedical research than to scientists that search for the latest findings in scientific papers with our limited capacities to process such information.

I think that S. Pelech above

I think that S. Pelech above brings up a very illustrative example of the benefits of data sharing, and why it should be encouraged whenever possible. Some of this may be partially unique to bioinformatics / computational-rich aspects of biology [as data sets are often easier to share than clinical specimens, for a number of reasons]. However, I wonder if another [at least minor] reason for the study results might be a genuine lack of time and the short-staffed nature of many labs these days in an era of reduced funding. I wish that I had the time to follow up on all requests for data or other information with more rich content than minimally required; however, as organizational budgets shrink, each employee's To Do list grows, and there are only so many hours in the day...