The Body: The Complete HIV/AIDS Resource
Follow Us Follow Us on Facebook Follow Us on Twitter 
Professionals >> Visit The Body PROThe Body en Espanol
  • Email Email
  • Glossary Glossary
ICAAC 2007 Study Summaries: Eric Stawiski

September 25, 2007

Listen (2MB, 5 min.)
Welcome. This is Bonnie Goldman, Editorial Director of The Body PRO. I'm in Chicago at ICAAC 2007, one of the year's HIV conferences. Right now, I'm in the poster session where researchers are standing in front of their posters. There's nothing like hearing the results of research directly from those who actually conducted the research. It is these women and men who are transforming HIV treatment and care. In this podcast, the researchers will introduce themselves and then summarize their study. After their summary, I'll ask a few questions.

My name is Eric Stawiski. I'm a senior bioinformatics scientist at Monogram Biosciences. Basically, my poster is about trying to predict coreceptor usage from the genotype. This is primarily done with the V3 loop, actually. So determining tropism will be relevant for the new class of drugs that are coming out, particularly maraviroc. The Trofile assay from Monogram was used to screen these patients, and is basically used as the standard because it correlates very well with the clinical outcome data.

So, what we were trying to do with this poster is use the genotype from roughly 600 clinical samples, just under 600. And basically there are some challenges in actually using the genotype to predict tropism. One is that the sequences themselves are very diverse, so there's a lot of sequence heterogeneity, and actually deciphering the chromatograms can be quite challenging, actually.

Because of that, there are actually a large number of possible combinations that you can have in the V3, because you can have more than one amino acid in any one particular position. So one strategy is to generate all the possible combinations out there. And we're basically using standard algorithms that are published. So these are basically state-of-the-art algorithms predicting tropism. One is the 1125 chargeable. One is a decision tree. One is a position-specific scoring matrix. And one is a support vector machine.

Basically, when we try to do these predictions, either on a single V3, or combinations of the V3, and maximizing the sensitivity by seeing if the algorithms predict X4 usage or R5 usage, it turns out that the sensitivity is quite low. That means that we're basically calling a lot of actual X4 viruses --we're saying that they use R5. So the sensitivities basically range from 24 percent to 56 percent. But the 56 percent actually comes at a trade off with a specificity, so it actually has the most total overall discordance. It has the highest total overall discordance.

Essentially, the take home message is that the algorithms aren't there yet for clinical utility, so you can't use just genotype to predict coreceptor usage, actually, at this point.

I see that it says by these estimates, two or three out of every four treatment-experienced individuals with X4 or mixed would be misidentified as R5 tropic?

That's right, yes. That's where the sensitivity comes in. By sensitivity, we mean the total number of correctly predicted X4 using viruses, divided by the total number of actual X4 viruses. So those were the numbers that would probably be really important, and those are the numbers that turn out to be quite low, actually.

So are there any recommendations for the future?

Yes. The Trofile assay, I think, is the standard right now, for Monogram. And it actually measures it experimentally in the lab through an actual assay.

So that's the best one, but it's not perfect.

Yes. Basically it's the one that Pfizer used for their clinical trials for maraviroc.

But they are still perfecting it, I noticed. There's another poster.

Yes. There's a next version two posters down.

Thank you very much.

  • Email Email
  • Printable Single-Page Print-Friendly
  • Glossary Glossary

This article was provided by The Body PRO. Copyright © Body Health Resources Corporation. All rights reserved.

Please note: Knowledge about HIV changes rapidly. Note the date of this summary's publication, and before treating patients or employing any therapies described in these materials, verify all information independently. If you are a patient, please consult a doctor or other medical professional before acting on any of the information presented in this summary. For a complete listing of our most recent conference coverage, click here.