Estimating HIV-1 Viremic Time From Reservoir Sequence Diversity With Uncertainty Quantification

Estimating HIV viremic time helps better understand reservoir dynamics and inform cure trials. Traditional approaches rely on serological assays or CD4 counts, which can lack precision. Sequence-based estimates using pretreatment RNA diversity help address this, but are limited by immediate ART initiation.

We developed Bayesian models to predict viremic time from diversity in HIV reservoir sequences in 2 cohorts from Uganda and Sweden. We computed 6 diversity metrics for gp41, RT, and matrix p17. We fitted 36 models per region using slope-fitting and weighting strategies and evaluated predictive accuracy and diagnostics. We validated top models in participants with diagnosis dates but unknown seroconversion dates.

Reservoir diversity increased with viremic time. Models using unique RT and gp41 sequences performed well; combining regions improved predictions, especially using nucleotide diversity and mean TN93 distances. In validation, estimates and credible intervals that aligned with diagnosis dates; and log-transformed sequence count weights increased precision. Models using matrix p17, complex metrics, or identical sequences showed weaker performance.

We present a new Bayesian approach to estimate HIV viremic time from reservoir sequences. This approach works across subtypes and chronic infection, uses simple diversity metrics, and may support research on reservoir dynamics and cure.