Fig. 4. Probability distributions and bNAb probability scores across cohorts.
a Pgen distribution as well as mean and median Pgen for heavy and light chains derived from uninfected control (CTRL, n = 57 individuals), HIV-1-infected (n = 34 individuals), and HCV-infected (n = 12 individuals) cohorts by IGoR. Averages and error bars show the mean ± SD for mean Pgen values and the median ± median absolute deviation (MAD) for median Pgen values. b PSHM distributions for heavy and light chains for the same cohorts as in a. Averages and error bars show the mean ± SD for mean PSHM values and the median ± median absolute deviation (MAD) for median PSHM values. Pgen and PSHM distributions in a and b show means as solid lines and SD as shaded areas. c Correlation plots of Pgen for neutralizing antibody heavy (n = 70), kappa (n = 33), and lambda chains (n = 26) derived from either uninfected control (x-axis) or chronically infected cohorts (HIV-1, HCV, y-axis). d Correlation plots of PSHM for the same chains as in c. e Correlation plot for bNAb heavy (n = 70) and light chain probability scores (n = 59) derived from HCV or HIV-1 cohorts (x-axis) in comparison to uninfected control (CTRL, y-axis). Cohort-specific Pgen and PSHM of bNAbs were used to calculate the bNAb probability scores S using coefficients from the linear regression on uninfected individuals (Fig. 2g). The dashed line represents identity. f Comparison as in e with the HIV-1 cohort stratified by antiretroviral therapy (on ART, n = 22 individuals; off ART, n = 12 individuals). g Comparison as in e with the HIV-1 cohort stratified by serum neutralization breadth against the global HIV-1 panel into high (≥66%, n = 8 individuals), intermediate (≥33%, n = 5 individuals), or low (>0%, n = 16 individuals) breadth and no neutralization (0%, n = 4 individuals) to recalculate Pgen and PSHM of bNAbs for these cohorts. Serum neutralization data for one individual was not available. ART anti-retroviral therapy. Source data are provided as a Source Data file.