Gordon, the superhero supercomputer
A big computer that does super things is the best way to describe our newest non-human collaborator, Gordon the Supercomputer. Housed at the San Diego Supercomputer Center at the University of California San Diego, Gordon and his human mastermind, collaborator Dr. Wayne Pfeiffer, have been working with us over the past few years to try to unravel the mystery of disease transmission in birds.
Avian mycobacteriosis is a rare, but important bacterial disease of birds caused by a bacterium similar to the one that causes human tuberculosis. Like human TB, it has an insidious onset and takes months or even years to manifest making it difficult to know who may be infected. The disease can be especially problematic when it halts our captive breeding conservation programs for species extinct in the wild, such as the Guam kingfisher.
To investigate the transmission potential of this disease, we are using powerful, state-of-the art methods. These involve whole-genome sequencing of bacteria collected from many infected birds followed by computational analysis to determine how closely related the genomes are. If bacteria from two birds have genomes that differ by only a few mutations, then one of those birds may have infected the other. On the other hand, if the genomes differ by many mutations, then transmission of bacteria between the host birds is highly unlikely.
Analyzing whole bacterial genomes requires the processing of large amounts of data in a timely manner. Each genome consists of about 5 million bases, i.e., letters in the four-letter DNA alphabet. Although this is not particularly large, the sequencer output was hundreds of times this size, or several gigabytes, to ensure that the entire genome was covered and to correct for occasional errors. Multiplying this by the 105 birds in our study resulted in hundreds of gigabytes of data to start with. These data ballooned to a few terabytes during the computational analysis. Overall, tens of thousands of core hours of computer time were consumed!
Analysis at this scale is still not practical on personal computers. However, Gordon the supercomputer made it easy, zooming through each genome in several hours time! With superhero capabilities, Gordon is helping to save birds in our care.
Photo by Alan Decker, UCSD
Our team at the Wildlife Disease Laboratories, Dr. Pfeiffer, and Gordon the supercomputer will soon have better answers as to how and when this disease is spread between birds.