Skip to main content
. 2022 Jan 18;17(1):e0262314. doi: 10.1371/journal.pone.0262314

Fig 4. Survey of correlation between sequence-based predictions and experimental measures of “fitness” based on replicative capacity.

Fig 4

Spearman correlation coefficients, ρ between prevalence-based measures of fitness as predicted by the Potts (blue) and independent (gray) models and experimental measurements related to replicative capacities are shown across four different HIV proteins: PR, RT, IN, and p24 CA. Experimental data are obtained from [3, 21, 28, 3941]. Experiments reporting fitness measurements for random mutations are marked with an “R” and experiments reporting drug-resistance only mutations are marked with a “D”. Correlation is not consistent between different experiments for the same protein. The Potts model generally (marginally) outperforms the independent model in capturing experimentally measured replicative capacities or measures related to replicative capacities.