Model 2A GA optimialisation.Rmd

```{r}
library(tidyverse)
library(GillespieSSA2)
library(GA)
library(doParallel)
```

```{r}
rm(list = ls()) # clear environment

# Define initial states
ini_state <- c(
  DD = 10, 
  p53 = 10, 
  p53P = 17,
  MDM2 = 10,
  p21 = 4
)
param_names <- c("CP_DD", "CD_DD", "CP_p53", "CD_p53", "CD_p53_MDM2", "phos", "dephos", "CD_p53P", "CD_p53P_MDM2", "CP_MDM2", "CD_MDM2", "CP_p21", "CD_p21")
  
# Define the new objective function based on p21 protein levels
objective_function <- function(params) {

  names(params) <- param_names

  
  reactions1 <- list(
    reaction(~params["CP_DD"], c(DD = +1), name = "Prod_DD"),
    reaction(~params["CD_DD"] * p53P, c(DD = -1), name = "Decay_DD"),

    reaction(~params["CP_p53"] * DD, c(p53 = +1), name = "Prod_p53"),
    reaction(~params["CD_p53"] * p53, c(p53 = -1), name = "Decay_p53"),
    reaction(~params["CD_p53_MDM2"] * p53 * MDM2, c(p53 = -1), name = "Inhibition_p53_MDM2"),
    
    reaction(~params["phos"] * p53 * DD, c(p53 = -1, p53P = +1), name = "Phos_p53"),
    reaction(~params["dephos"] * p53P, c(p53P = -1, p53 = +1), name = "Dephos_p53P"),
    reaction(~params["CD_p53P"] * p53P, c(p53P = -1), name = "Decay_p53P"),
    reaction(~params["CD_p53P_MDM2"] * p53P * MDM2, c(p53P = -1), name = "Inhibition_p53P_MDM2"),
    
    reaction(~params["CP_MDM2"] * p53, c(MDM2 = +1), name = "Prod_MDM2"),
    reaction(~params["CD_MDM2"] * DD * MDM2, c(MDM2 = -1), name = "Decay_MDM2"),
    
    reaction(~params["CP_p21"] * p21_mRNA, c(p21 = +1), name = "Prod_p21"),
    reaction(~params["CD_p21"] * p21, c(p21 = -1), name = "Decay_p21")
  )
  
  out <- ssa(
      initial_state = ini_state,
      reactions = reactions1,
      params = params,
      method = ssa_exact(),
      final_time = 24,
      census_interval = .01,
      verbose = FALSE
    )
  
  # Extract p21 data
  p21_data <- out %>% filter(variable == "p21")
  
  # Calculate the objective metric based on desired behavior
  oscillation_metric <- sd(p21_data$value)
  trend_metric <- mean(diff(p21_data$value))
  combined_metric <- oscillation_metric + abs(trend_metric)
  
  return(-combined_metric)
}
```

```{r}
# Define parameter bounds
lower_bounds <- c(0.001, 0.001, 0.001, 0.001, 0.001, 0.001, 0.001, 0.001, 0.001, 0.001, 0.001)
upper_bounds <- c(10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10)

# Run the Genetic Algorithm optimization
ga_result <- ga(
  type = "real-valued",
  fitness = objective_function,
  lower = lower_bounds,
  upper = upper_bounds,
  popSize = 100,
  maxiter = 100,
  run = 100,
  keepBest = TRUE,
  parallel = TRUE,
  seed = 123
)

# Extract the best parameter set
best_params <- ga_result@solution

# Print the best parameters
print(best_params)
```