The fight against spam help in the search for HIV vaccine

Help solve some of the most pressing global challenges is an important point on the agenda of Microsoft Research (MSR) - and through their experience, knowledge, partnerships and the power of software that can help solve them. In the case of finding a vaccine for HIV, it is the application of high computing power, combating HIV through data and, surprisingly, with the experience of Microsoft Research to build spam filters for find a solution.

More than 1.8 million people die from HIV related causes each year - about 5 thousand deaths a day. One of the biggest challenges in the fight against HIV is the virus constantly mutates to avoid attacks by the immune system - may change in an infected person has changed the flu virus in its history. This becomes extremely difficult to accurately analyze the virus and develop therapies that target their weaknesses elusive. Each mutation means another variable to identify and understand. To complicate matters further, the immune response of each individual varies significantly, the immune system of some people can fight the virus, allowing them to live for years without any treatment, while others are sick more fast because your body can not withstand the attack of the virus.

As with many Microsoft Research projects, this work is carried out in conjunction with many other experts in the field. Tests of a vaccine in Durban, South Africa, are led by Bruce Walker of the Ragon Institute of Massachusetts General Hospital, MIT and Harvard, and a professor of medicine at the University of KwaZulu-Natal . They met through the Center for AIDS Research Program in South Africa and the Research Institute for Tuberculosis and HIV in KwaZulu-Natal. This test program generates large amounts of data, whose analysis is a major challenge.

That's where David Heckerman and Jonathan Carlson of Microsoft Research, together with a tool called the Microsoft Computational Biology PhyloD Tool into action. This software allows for efficient data mining that later leads to an analysis of each cell that helps to detail the patterns of virus for further analysis. PhyloD has algorithms, codes and visualization tools that perform complex pattern recognition and analysis - enabling Heckerman and colleagues to learn how different immune systems respond to changes in the virus.