Non-profit collaboration pioneers new resistance interpretation system

This article is more than 21 years old.

A new system for predicting treatment response from HIV drug resistance mutations has been launched this week, by a non-profit collaborative group of scientists in Europe and North America.

The HIV Response Database Initiative will collect genotype and treatment response information in order to develop a complex model of the relationship between HIV genotype (the pattern of resistance mutations) and response to treatment.

At present, prediction of response to treatment is achieved by means of an algorithm, developed by matching data on treatment response and genotype. The algorithm predicts whether a specific drug will be active in a patient with a particular mutation or pattern of mutations. This approach is most likely to be accurate when many patients in the database used to create the algorithm share a common genotypic pattern. If a patient has an unusual resistance pattern, the database may not contain enough information to predict accurately the outcome of treatment with a specific drug.

Glossary

anonymised data

Information about a patient from which the name, address and other identifying information has been removed.

drug resistance

A drug-resistant HIV strain is one which is less susceptible to the effects of one or more anti-HIV drugs because of an accumulation of HIV mutations in its genotype. Resistance can be the result of a poor adherence to treatment or of transmission of an already resistant virus.

matched

In a case-control study, a process to make the cases and the controls comparable with respect to extraneous factors. For example, each case is matched individually with a control subject on variables such as age, sex and HIV status. 

genetics

The science of inheritance: the study of how genes are passed down throughout generations, as well as the study of individual genes and how they affect the body.

In the past two years, Dr Brendan Larder and colleagues, formerly of Virco and Visible Genetics, have become interested in the use of neural networking by computers to try and develop more accurate methods of predicting response according to genotype.

Neural networking runs many thousands of calculations to test the accuracy of particular formulae for predicting response according to genotype, all the time adjusting the degree of importance given to particular mutations in order to arrive at the `best fit`. The computer also uses the learning process in itself to refine the way it makes decisions about future weighting of particular mutations. The accuracy of the technique is vastly improved by larger quantities of data, and the researchers are now appealing for institutions around the world to contribute anonymised patient data to the database.

"The key to success of this initiative is the amount of clinical data in the database", commented Dr Brendan Larder, Chair of the RDI Scientific Core Group. "While no one company or organisation can generate enough on its own, the RDI represents a unique opportunity for individuals institutions and companies around the world to share their data and I am delighted that it is gathering such momentum."

The HIV Resistance Response Database Initiative is a non-profit organisation established by Brendan Larder. Members of the Scientific Core Group include Julio Montaner and Richard Harrigan of the British Columbia Centre for Excellence in HIV/AIDS, Victor DeGruttola of the Harvard School of Public Health, Scott Wegner of the US Military HIV Research Group and Maurizio Zazzi of the University of Siena, Italy.

The group’s new website will offer clinicians the opportunity to submit queries for free advice, and eventually, to interrogate the database over the internet using a menu of query options. For example, a researcher may want to study the response of all patients with an M184V mutation. Or they may want to compare response to AZT for all patients with an M184V plus or minus any of the signature AZT mutations.

The group is also looking for research collaborators, both academic and commercial.