posted on Feb, 27 2005 @ 10:04 AM
Machine learning algorithms commonly used to compress digital images and recognise patterns in email spam might also be able to help scientists find
an effective vaccine for HIV.
The algorithms, developed by US software company Microsoft, are being adapted for vaccine development in collaboration with researchers from the
University of Washington in Seattle, US, and Royal Perth Hospital in Western Australia.
Prototype vaccines, produced using the algorithms, are tested against immune cells taken from HIV-infected patients. The trial vaccines are based on a
"cell-mediated" response - meaning they activate the body's T-cell system to attack viral cells.
Andrew McMichael, an HIV vaccine expert at the University of Oxford, UK, agrees the approach could eventually help to counteract several strains of
HIV simultaneously. But he says important problems must first be overcome.
Please visit the link provided for the complete story.
This is too funny, Microsoft to the rescue!
HIV mutates rapidly, thus evading the human immune system. This means that vaccines developed to counteract one strain may not be effective against
But the researchers hope that algorithms capable of finding patterns in digital information could also help identify key genetic features across many
different strains of HIV. This could enable them to engineer an HIV vaccine that is effective against several strains at once.
"This is the first attempt to build the variability of HIV into a vaccine," says James Mullins at the University of Washington, where the approach
is undergoing laboratory testing.
[edit on 27-2-2005 by Bourgeoisie]