Skip to main content
. 2019 Nov 28;221(7):1135–1145. doi: 10.1093/infdis/jiz563

Figure 4.

Figure 4.

Immunological and clinical variables that relate to total human immunodeficiency virus (HIV) DNA. A, Boosted regression trees model to assess predictors of baseline total HIV DNA (49 predictors; n = 60); box plots show the summary of 100 model runs. Influential predictors were defined as those whose relative contribution was >100, divided by the total number of covariates; this value is indicated by dashed vertical line. Asterisks denote variables that are also selected in B. B, Least absolute shrinkage and selection operator (LASSO) output for predictors of baseline total HIV DNA (49 predictors; n = 60; deviance explained, 0.62). Variables that do not significantly contribute to the model have a coefficient of 0; only those with a nonzero coefficient are shown, and missing values were imputed. Coefficients represent the change in log10 total HIV DNA per 1% higher predictor variable. C, Schematic illustrating which factors were associated with HIV DNA level during treated primary HIV infection in this analysis. Abbreviations: ART, antiretroviral therapy; CM, central memory; EM, effector memory; EMRA, effector memory T cells re-expressing CD45RA; PD-1, programmed cell death protein 1; sPD-1, soluble PD-1; sTim-3, soluble Tim-3; TIGIT, T cell immunoreceptor with immunoglobulin and ITIM domains; Tim-3, T cell immunoglobulin and mucin-domain containing protein 3; TM, transitional memory; VL, viral load.