Mystery Rays
If you like anything I put up on this blog, you'll definitely like Mystery Rays. Dr. Ian York does a great job of presenting new and interesting aspects of immunology/virology. Topics range from virus evolution to cancer and tend to center around immune escape. In addition to good reading, I have also used this site to find background information and paper suggestions. Mostly, this is because Dr. York covers some of the stuff they don't mention much in introductory immunology courses (for example, immunodominance). And, he does it in a clear and informative manner.
He has a new post up concerning the possibility of HIV actually increasing its virulence.
So, if that sparked your interest, go check it out!
Mystery Rays from Outer Space
http://www.iayork.com/MysteryRays/2009/04/16/is-hiv-becoming-more-virulent/
Friday, April 17, 2009
Saturday, April 4, 2009
System.out.println("The answer is 42");
The Dawning Age of Robot Scientists
Introduction
Friday has become my favorite day. Not because it's the weekend, oh no. No, Fridays are my favorite because I'm such a geek that I'm excited when The Journals come out. And even if you don't regularly scan their table of contents, you'd be hard pressed to miss the headlines in the science sections from the general news media: The first steps to a robot scientist are here people!
(All on the same story, but from different sources)
http://www.nytimes.com/2009/04/07/science/07robot.html
http://news.bbc.co.uk/2/hi/science/nature/7979113.stm
http://sciencenow.sciencemag.org/cgi/content/full/2009/402/1
http://blog.wired.com/wiredscience/2009/04/robotscientist.html
The actual scientific paper in Science is called The Automation of Science by King et al. at Aberystwyth University. There is also a comment on it and a related paper in the same issue (though the news outlets do a good job as well.)
Essentially, this robot, termed Adam, is able to complete the scientific process. Whereas computers and automation today are capable of collecting vast amounts of data, they have so far been unable to process it. Human scientists usually need to look at the data and determine what's useful. However, King et al. designed a system that can not only interpret the data and form a hypothesis about it, but can also perform followup experiments to test this hypothesis. To do this, they designed and implemented several different programs, ranging from hardware interaction to data collection and data analysis.
Just a note on the hardware: Adam is pretty beautiful, consising of multiple fridges, incubators, and robot arms. And, it never gets tired. The experiments involved growth curves of various yeast strains, which Adam can do hundreds of times a day. Just thinking about a human doing that amount of pipetting gives me thumb arthritis.
What Did Adam Actually Do?
Being a machine, Adam is great at hundreds of simple tasks and measurements. After taking specific yeast strains from the freezer, Adam can inoculate them in media and measure the ODs at various time points to establish a growth curve.
In order to test Adam, the authors decided to look at orphan enzymes, or enzymes for which the corresponding gene(s) haven't been identified. These enzymes have been well characterized in terms of their biochemical actions, so Adam is able to screen for enzyme activity by recording growth curves in specific media. Adam was also able to take advantage of yeast knockout strains for which the gene was known. Thus, Adam is able to take yeast that lack a gene and see if it is still able to grow when tested in a variety of limiting medias supplemented with certain metabolites. From this data, Adam is able hypothesize about the unknown gene encoding the orphan enzyme and is able to run further tests to confirm this hypothesis.
But Wait, There's More! (Bioinformatics)
One of the software packages I didn't talk about was their relational database of proteins and genes known to be present in yeast. Although the authors aren't descriptive, this is probably something similar to the DAVID Bioinformatics database (http://david.abcc.ncifcrf.gov/), Gene Ontology(http://www.geneontology.org/), or Pathways Analysis(http://www.ingenuity.com/products/pathways_analysis.html%20to)-- to name a few. I think this is the most important part of Adam, its "memory". Without such databases, it would be hard to understand gene and protein interaction and formulate an educated hypothesis.
Conclusion
Mass data collection, expanding bioinformatics tools, increasingly sophisticated programming, and efficient hardware are converging at the biological sciences, extending capability and knowledge at an exponential pace. Adam is only the beginning.
(If you don't believe me, read Ray Kurzweils The Singularity is Near, which, though on a slightly different topic, is on the same general trend. Or read about Eve at the group’s website, below.)
The paper is:
King et al. The Automation of Science. Science, vol. 324, April 3 2009.
http://www.ncbi.nlm.nih.gov/pubmed/19342587
The group's website can be found at:
http://www.aber.ac.uk/compsci/Research/bio/robotsci/
There are videos of Adam in action and other interesting things on the site
Introduction
Friday has become my favorite day. Not because it's the weekend, oh no. No, Fridays are my favorite because I'm such a geek that I'm excited when The Journals come out. And even if you don't regularly scan their table of contents, you'd be hard pressed to miss the headlines in the science sections from the general news media: The first steps to a robot scientist are here people!
(All on the same story, but from different sources)
http://www.nytimes.com/2009/04/07/science/07robot.html
http://news.bbc.co.uk/2/hi/science/nature/7979113.stm
http://sciencenow.sciencemag.org/cgi/content/full/2009/402/1
http://blog.wired.com/wiredscience/2009/04/robotscientist.html
The actual scientific paper in Science is called The Automation of Science by King et al. at Aberystwyth University. There is also a comment on it and a related paper in the same issue (though the news outlets do a good job as well.)
Essentially, this robot, termed Adam, is able to complete the scientific process. Whereas computers and automation today are capable of collecting vast amounts of data, they have so far been unable to process it. Human scientists usually need to look at the data and determine what's useful. However, King et al. designed a system that can not only interpret the data and form a hypothesis about it, but can also perform followup experiments to test this hypothesis. To do this, they designed and implemented several different programs, ranging from hardware interaction to data collection and data analysis.
Just a note on the hardware: Adam is pretty beautiful, consising of multiple fridges, incubators, and robot arms. And, it never gets tired. The experiments involved growth curves of various yeast strains, which Adam can do hundreds of times a day. Just thinking about a human doing that amount of pipetting gives me thumb arthritis.
What Did Adam Actually Do?
Being a machine, Adam is great at hundreds of simple tasks and measurements. After taking specific yeast strains from the freezer, Adam can inoculate them in media and measure the ODs at various time points to establish a growth curve.
In order to test Adam, the authors decided to look at orphan enzymes, or enzymes for which the corresponding gene(s) haven't been identified. These enzymes have been well characterized in terms of their biochemical actions, so Adam is able to screen for enzyme activity by recording growth curves in specific media. Adam was also able to take advantage of yeast knockout strains for which the gene was known. Thus, Adam is able to take yeast that lack a gene and see if it is still able to grow when tested in a variety of limiting medias supplemented with certain metabolites. From this data, Adam is able hypothesize about the unknown gene encoding the orphan enzyme and is able to run further tests to confirm this hypothesis.
But Wait, There's More! (Bioinformatics)
One of the software packages I didn't talk about was their relational database of proteins and genes known to be present in yeast. Although the authors aren't descriptive, this is probably something similar to the DAVID Bioinformatics database (http://david.abcc.ncifcrf.gov/), Gene Ontology(http://www.geneontology.org/), or Pathways Analysis(http://www.ingenuity.com/products/pathways_analysis.html%20to)-- to name a few. I think this is the most important part of Adam, its "memory". Without such databases, it would be hard to understand gene and protein interaction and formulate an educated hypothesis.
Conclusion
Mass data collection, expanding bioinformatics tools, increasingly sophisticated programming, and efficient hardware are converging at the biological sciences, extending capability and knowledge at an exponential pace. Adam is only the beginning.
(If you don't believe me, read Ray Kurzweils The Singularity is Near, which, though on a slightly different topic, is on the same general trend. Or read about Eve at the group’s website, below.)
The paper is:
King et al. The Automation of Science. Science, vol. 324, April 3 2009.
http://www.ncbi.nlm.nih.gov/pubmed/19342587
The group's website can be found at:
http://www.aber.ac.uk/compsci/Research/bio/robotsci/
There are videos of Adam in action and other interesting things on the site
Subscribe to:
Posts (Atom)