Mathematical model for HHV8 viral spread
Sub-directories are:
include - header files
src - C++ source files
Linux - build/run directory (Linux-based)
OSX - build/run directory (Mac-based)
scripts - shell and perl scripts for launching runs, scoring fits, etc.
data - PTID-specific criteria files (used for fitting)
inputs - input files and PTID lists (all and by ARM)
The simulation executable, hhv8_sim, can be built and run in either the Linux or OSX (Mac) sub-directory. It is designed to be run with an input file that specifies parameter values or a distribution (mean & stddev) for them. Originally, the simulation outputs were scored against values from a criteria file containing summary statistics from multiple participants using binned levels of shedding, episode peaks, etc. The scores were written to a log file for post-processing into score files (csv) by the appropriate perl script from the scripts sub-directory.
To score runs against data from an individual, the simulation must be run with parameters drawn for that individual and the output of the simuation saved for special post-processing by a separate script (indiv_rollup.pl) launched in conjunction with that participant's criteria file. The criteria file in this case contains shedding percentage, mean and peak log10 viral levels for that individual from the 28-day observation period. All criteria files are contained in the data sub-directory.
Several "models" were evaluated when fitting to participant data. A model in this case is defined as a set of parameters to be fitted and their associated mechanistic effects. The best model from our testing was one that varied infectivity of a cell (beta), the death rate of productively infected cells (an), the killing rate of each T cell (fpos) and the reactivation rate from latency (latent_inf). It is identified as model 1 in the scripts. Two alternative models can be fitted using the scripts. They are model 2, which assigns infectivity to virions, and model 3 which does the same and allows spread across regions via extracellular virus.