Modeling HIV Protease Drug Resistance
Author Information
Author(s): Lapins Maris, Eklund Martin, Spjuth Ola, Prusis Peteris, Wikberg Jarl ES
Primary Institution: Uppsala University
Hypothesis
Can proteochemometrics effectively predict the susceptibility of mutated HIV strains to protease inhibitors?
Conclusion
The proteochemometric model can reliably predict HIV protease susceptibility to various inhibitors based on viral genotype.
Supporting Evidence
- The model achieved a predictive ability of Q2 = 0.87.
- External validation showed that the model could predict susceptibilities for omitted inhibitors with Q2 inhibitors = 0.72.
- Model-3, which included intra-protease cross-terms, demonstrated the best performance.
- Statistical validity was confirmed through permutation testing.
Takeaway
This study created a model that helps doctors choose the right medicine for HIV by predicting how well the virus will respond to different drugs.
Methodology
The study used proteochemometrics to analyze the susceptibility of HIV protease variants to seven protease inhibitors, correlating physico-chemical properties with susceptibility data.
Potential Biases
Potential biases may arise from the reliance on existing susceptibility data and the specific inhibitors studied.
Limitations
The model's predictions are limited to the inhibitors included in the study and may not account for all possible mutations.
Participant Demographics
The study analyzed 828 unique HIV protease variants.
Statistical Information
P-Value
0.72
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website