Sharing Data

by techtyper on July 1, 2009

In BMC Research Notes, Vincent S. Smith of the Natural History Museum in London published Data Publication: Towards a Database of Everything. As Smith wrote:

… outside a handful of disciplines, publication of science data is the exception, not the rule.

As one driver for increasing the publication of data, Smith suggested:

If data publication is to become a part of normal scientific practice it has to be easy to achieve.

That might be so. If it is, how could it be made easier? On the other hand, easier data publishing could lead to a decline in the value of the data that do get published. So if it becomes incredibly easy to publish data, what will maintain data quality?

For some fields, especially genomics and personalized medicine, data would also be more useful if it were easier to share. Recently, I wrote about this topic in Sharing the Wealth of Data in Scientific American Worldview. In medicine, especially on the pharmaceutical side, much of the data could always remain proprietary. Moreover, some patients will want their data private as well, or at least protected.

So although issues of data publication and sharing both depend on technology, a range of other issues—economic, political, sociological, and so on—must be considered, as well.

Share this with your colleagues:
  • Digg
  • Facebook
  • Google Bookmarks
  • LinkedIn
  • Technorati
  • connotea
  • del.icio.us

{ 0 comments }

Mining Genomes for New Drugs

by techtyper on June 23, 2009

The June 16, 2009, BMC Bioinformatics includes Automated Genome Mining for Natural Products by Michael HT Li and colleagues at the University of Michigan. This article describes a program called NP.search, which one of the authors—David H. Sherman—describes by saying:

This software tool mines microbial genomes for clusters of secondary metabolite biosynthetic genes, decodes the enzymatic machinery, and translates this information into predicted 2- or 3-D natural product chemical structures.

So rather than screening organisms—in wet labs—for what they make and what might be useful, this program—open-source code, written in C++—can search a microorganism’s genome for what it can make, and then build structural models of the natural products. Eventually, such models could be tested computationally against disease targets. So far, though, some fine tuning remains. For example, the authors wrote:

Currently, the best predicted structure of the molecule differs significantly in many cases from the actual molecule because of non-functional domains and unrecognized post-tailoring modifications.

Still, this approach could enhance the future approach to drug discovery by making it possible to “screen” far more compounds—doing so more quickly and economically.

Share this with your colleagues:
  • Digg
  • Facebook
  • Google Bookmarks
  • LinkedIn
  • Technorati
  • connotea
  • del.icio.us

{ 0 comments }

Managing Genomics Data

June 19, 2009

Everyone in the life sciences knows that data acquisition keeps bringing in far more data, but few people know how to handle it. When it comes to genomics data, part of the solution involves the right hardware, but researchers also need rules for how hardware and software work together to keep track of the data. [...]

Share this with your colleagues:
  • Digg
  • Facebook
  • Google Bookmarks
  • LinkedIn
  • Technorati
  • connotea
  • del.icio.us
Read the full article →

Doherty and Turner on Influenza Basics

June 17, 2009

In the May 2009 Journal of Biology, Nobel laureate Peter C. Doherty and Stephen J. Turner published Q&A: What Do We Know about Influenza and What Can We Do about it? This article covers the molecular basics behind an influenza pandemic and the basic approach behind developing an influenza vaccine. Doherty and Turner also indicate some of [...]

Share this with your colleagues:
  • Digg
  • Facebook
  • Google Bookmarks
  • LinkedIn
  • Technorati
  • connotea
  • del.icio.us
Read the full article →

The Computer Challenge of Personalized Medicine

June 13, 2009

In the past few years of hearing about personalized medicine, I have usually been told about molecular hurdles that needed clearing. For example, genotyping needed to be less expensive. Today, though, I hear more about the the challenge of using the data, rather than just getting them. In the May 2009 issue of Genome Medicine, [...]

Share this with your colleagues:
  • Digg
  • Facebook
  • Google Bookmarks
  • LinkedIn
  • Technorati
  • connotea
  • del.icio.us
Read the full article →