Evaluation of the HIV-1 Polymerase Gene Sequence Diversity for Prediction of Recent HIV-1 Infections Using Shannon Entropy Analysis

HIV-1 incidence is an important parameter for assessing the impact of HIV-1 interventions. The aim of this study was to evaluate HIV-1 polymerase (pol) gene sequence diversity for the prediction of recent HIV-1 infections. Complete pol Sanger sequences obtained from 45 participants confirmed to have recent or chronic HIV-1 infection were used. Shannon entropy was calculated for amino acid (aa) sequences for the entire pol and for sliding windows consisting of 50 aa each. Entropy scores for the complete HIV-1 pol were significantly higher in chronic compared to recent HIV-1 infections (p < 0.0001) and the same pattern was observed for some sliding windows (p-values ranging from 0.011 to <0.001), leading to the identification of some aa mutations that could discriminate between recent and chronic infection. Different aa mutation groups were assessed for predicting recent infection and their performance ranged from 64.3% to 100% but had a high false recency rate (FRR), which was decreased to 19.4% when another amino acid mutation (M456) was included in the analysis. The pol-based molecular method identified in this study would not be ideal for use on its own due to high FRR; however, this method could be considered for complementing existing serological assays to further reduce FRR.

Automated summary

This study evaluated HIV-1 polymerase gene sequence diversity for predicting recent HIV-1 infections. The results showed that the entropy scores for the HIV-1 pol were higher in chronic infection compared to recent infection, and certain amino acid mutations could differentiate between recent and chronic infections. Different amino acid mutation groups had varying performance in predicting recent infections, and the false recency rate was reduced when an additional mutation (M456) was included.