Investigating the determinants of therapy response in HIV infected patients using a data mining approach, with special focus on drug resistance and

1 January, 2003 to 31 December, 2004


  • Jan Ramon

The project aims at investigating which factors determine therapy response in HIV infected patients, using a data mining approach. In the European context, the B subtype of HIV is most common, hence most expertise on HIV concerns the B subtype. However, Belgium is special as it has a high prevalence of non-B subtypes. Subtype C in particular is starting to circulate amongst Belgian patients. This is also the subtype most common in South-Africa. The aim of this collaborative effort is to build an extended database of resistance genotypes and therapy response data for non-B subtypes (using Belgian and South-African data), with a special focus on subtype C, and to study these data using data mining and other techniques. Particular attention will be devoted to evaluating the current (mainly subtype-B based) resistance scoring techniques, improving them and/or developing new ones, building algorithms for predicting therapy response from genotypic data, and setting up an automated subtyping tool.