Posted by
George Faulkner in
We try on this blog to stay fairly grounded in technological topics that are affecting the way we live and work today. Theorizing too much on what might be possible in 20 years can sometimes get in the way of what we can actually accomplish today. That said, this is worth the digression. IBM Research announced a fascinating project today with significant implications in how we conduct personalized medicine in the future. I’ll let the researchers behind the science tell the story themselves in the video above. But I’ll posit this one question. Considering what only a few years ago cost $3 billion might soon cost only $1000, what role could inexpensive individual genome sequencing have on personalized medicine?
(Stay tuned in the near future for more on this topic – in particular, the how privacy needs to play a central role in these personal technological advances.)
Previous post
Next post
1:07 pm
IBM DNA Transistor…
Your personal DNA sequencing co-processor is not far away from now. IBM taking big leaps to make it happen.
This technology will totally reshape the landscape of medicine and how drugs will be prescribed in the near future.
……
Posted by: Better than real
4:39 am
[...] For an ongoing conversation on this research project, visit the Smarter Planet blog: http://asmarterplanet.com/blog/2009/10/dna-transistor.html [...]
Posted by: IBM Research aims to build Nanoscale DNA Sequencer to help drive down cost of Personalized Genetic Analysis | Health Informer
12:48 am
[...] View video: View animation: For an ongoing conversation on this research project, visit the Smarter Planet blog: http://asmarterplanet.com/blog/2009/10/dna-transistor.html [...]
Posted by: IBM Research Aims to Build Nanoscale DNA Sequencer to Help Drive Down Cost of Personalized Genetic Analysis | Free Press Release
4:27 pm
Wow. Would love to see the raw signal data from this and train an ensemble of MLP artificial neural nets to predict base sequence under different conditions of voltage, current, solvent, buffer, conductivity etc. Even if there isn’t a global solution for the best possible prediction, there’s probably a pretty good local one in or near the accessible experimental data space, or a gradient of performance to suggest a direction to move in.
Crossing my fingers that the conditions could be optimized in real time to tune the detector along the continuum between random and highly-repetitive GC-rich regions.
Sounds like fun! Good luck!
Posted by: Boyd Steere