The PLoS Medicine Editors, Why Bigger Is Not Yet Better: The Problems with Huge Datasets, PLoS Medicine, February 2005. Excerpt: ' "Publishing results in traditional paper based way in a journal hides too much information." This is the verdict of Markus Ruschhaupt and colleagues who, in a paper in Statistical Applications in Genetics and Molecular Biology (3: article 37), discuss a paradigm for the presentation of complex data—in this case, from microarray analyses. The title of the article, “A Compendium to Ensure Reproducibility in High-Dimensional Classification Tasks,” may not lend itself easily to a clinical audience, but the underlying message to clinicians could not be more important: that, currently, studies involving large datasets, especially ones that have a clinical outcome, are so poorly reported (or possibly so poorly done) that many are not reproducible....So how can we ensure that the wealth of data pouring out of microarray and other molecular diagnostic studies is turned into meaningful knowledge? The Microarray Gene Expression Data Society has proposed a set of guidelines (MIAME) for the reporting of microarray data, and that all such data should be deposited in public databases. But as Ruschhaupt and others have shown, disclosure of results and data is not enough, since there is little consensus on the appropriate statistical analyses and many are developed on a case by case basis, which may not be reproducible, even by the authors....An ultimate aim for reporting would be the type of compendium discussed by Ruschhaupt and colleagues --"an interactive document that bundles primary data, statistical processing methods, figures, and derived data together with the textual documentation and conclusions." '
Posted by
Peter Suber at 3/04/2005 10:04:00 AM.
The open access movement:
Putting peer-reviewed scientific and scholarly literature
on the internet. Making it available free of charge and
free of most copyright and licensing restrictions.
Removing the barriers to serious research.
