BioMatters is a world leader in DNA sequencing, gene assembling and tree forming.
Exporting to 29 Universities world-wide.
geneious
My list of interesting things to do with computers:
1) Search and Match DNA sequences: Craig Venter has a big DECAlpha farm to do the HGSequence.
- But turns out that I could have stayed in NZ to join a world class DNA outfit
2) Simulate explosions for the military
3) Simulate explosions for CGI movies, with Orcs running about.
2) apparently is fairly similar to 3) which had the USA DOD concerned when Peter Jackson assembled the biggest Intel Farm in Wellington for LOTR)
I left Wellington to travel the world, with CGI in mind. Who knew Wgtn would be the place?
3) handle large integers to enable secure transactions and identification
5) simulate the climate crash
6) manage your Audio and Video info, run the Internet, telescopes, traffic lights etc.
7) run businesses, ie databases
I lost interest in CGI (endless noisy orc/goblin battles) so settled for 3)
1) would have been the most fantastic career choice.
DNA searching is at first glance simple. Do a Boyer-Moore on sequences encoded at 2 bits per Base (there are 4 bases)
- I hear that Boyer-Moore has been improved since I implemented it in 16-bit Intel Assembler in the 90's
DNA has peculiarities: Some Base changes dont change the Amino Acid, so thats another layer of search.
Then theres functional search: Some AminoAcids can be replaced with no change in protein function - eg some may simply be 'hydrophilic' to insert in membrane.
Then there are enzyme 'active site' vs backbones, which may be swapped?
Then there are mutations: a single Base loss means a 'Frame Shift' so that all downstream codons are off.
Combine with translocations and jumping genes, some limited back mutation and metres of ancient ERV remnants and the game gets interesting.
Nature, in the form of the MHC and the immune system, hunts for Aa sequences about 5 Aas long thats 20^5 = 1 in 3 million.
Any less and the VanderW forces wouldnt hold, so that may be as important as the odds of uniqueness.
I dont know how many DNA bases in a typical Repressor/Activator sequence. Obviously activator proteins must act on more than one gene, if a gene is required to produce an activation.
This is where the most worthwhile computer stuff is. Its just so damned hard to do Biochem. Life molecules have a tendency to behave badly in the undergraduate lab.
__________________________________________________________________
NewsFlash: Low-Cost NFC in Turkey:
"another Huawei-made device has been spotted, this time in Turkey. Called Sonic (aka Huawei T20), it's a mid-range device with NFC capability and pre-loaded Gingerbread. Turkish Turkcell is offering it for 150 EUR" (NZD255)
engadget
