pExit, param.pExit and timeDep.pExit

• pExit is the first required fundamental parameter and provides a daily probability for a host to leave the simulation (either cured, died, etc.).
• param.pExit is the list of functions needed to individually parameterize pExit (see Get started). The name of each function in the list has to match the name of the parameter it is sampling for pExit.
• timeDep.pExit allows for pExit to be dependent on the “absolute” time of the simulation, to account - for example - for seasonality or other external time-related covariates. By default, timeDep.pExit is set to FALSE.

nContact, param.nContact and timeDep.nContact

• nContact represents the number (expressed as a positive integer) of potentially infectious contacts an infected hosts can encounter per unit of time. At each time point, a number of contacts will be determined for each active host in the simulation. The number of contacts (i.e. the output of your function) has to be an integer and can be set to zero.
• param.nContact is the list of functions needed to individually parameterize nContact (see Get started). The name of each function in the list has to match the name of the parameter it is sampling for nContact.
• timeDep.nContact allows for nContact to be dependent on the “absolute” time of the simulation, to account - for example - for seasonality or other external time-related covariates. By default, timeDep.nContact is set to FALSE.

pTrans, param.pTrans and timeDep.pTrans

• pTrans is the heart of the transmission process and represents the probability of transmission over time (when a contact occurs).
• param.pTrans is the list of functions needed to individually parameterize pTrans (see Get started). The name of each function in the list has to match the name of the parameter it is sampling for pTrans.
• timeDep.pTrans allows for pTrans to be dependent on the “absolute” time of the simulation, to account - for example - for seasonality or other external time-related covariates. By default, timeDep.pTrans is set to FALSE.

pExit

For pExit, we choose a constant value, namely 0.08, i.e. an infected host has 8% chance to leave the simulation at each unit of time:

p_Exit_fct  <- function(t){return(0.08)}

Remember that pExit, like the other core functions has to be function of t, even if t is not used. Since pExit is constant here, there is no use for the “absolute” time of the simulation nor for the individual-based parameters. So param.pExit=NA, and timeDep.pExit=FALSE.

nContact

For nContact, we choose a constant function that will draw a value from a normal distribution with mean = 0.5 and sd = 1, round it, and take its absolute value:

n_contact_fct = function(t){abs(round(rnorm(1, 0.5, 1), 0))}

The distribution of nContact looks as follows:

At each time step and for each infected host, nContact will be drawn anew. Remember that nContact, like the other core functions has to be function of t, even if t is not used. Since nContact is constant here, there is no use for the “absolute” time of the simulation nor for the individual-based parameters. So param.nContact=NA, and timeDep.nContact=FALSE.

pTrans

We choose pTrans in the form of a threshold function: before a certain amount of time since initial infection, the host does not transmit (incubation time, which we call t_incub), and after that time it will transmit with a certain (constant) probability (which we call p_max). This function is dependent on the time since the host’s infection t:

p_Trans_fct <- function(t, p_max, t_incub){
    if(t < t_incub){p=0}
    if(t >= t_incub){p=p_max}
    return(p)
}

Because each host is different (slightly different biotic and abiotic factors), you can expect each host to exhibit differences in the dynamics of infection, and hence the probability of transmission over time. Thus, t_incub and p_max will be sampled for each host individually according to a certain distribution. t_incub will be sampled from a normal distribution of mean = 7 and sd = 1, while p_max will be sampled from a beta distribution with shape parameters α = 5 and β = 2:

t_incub_fct <- function(x){rnorm(x,mean = 7,sd=1)}
p_max_fct <- function(x){rbeta(x,shape1 = 5,shape2=2)}

Note that here t_incub and p_max are functions of x and not t (they are not core functions but individual-based parameters), and x enters the function as the number of draws to make.

Taken together, the profile for pTrans for a subset of 200 individuals in the population will look as follows:

pTrans is not dependent on the “absolute” time of the simulation, so timeDep.pTrans=FALSE. However, since we make use of individual-based parameters, we have to provide a param.pTrans as a list of functions. The name of each element within this list should have the same name that the core function (here pTrans) uses as argument, e.g.:

t_incub_fct <- function(x){rnorm(x,mean = 7,sd=1)}
p_max_fct <- function(x){rbeta(x,shape1 = 5,shape2=2)}

param_pTrans = list(p_max=p_max_fct, t_incub=t_incub_fct)

Running

Once nosoiSim is set up, you can run the simulation (here the “seed” ensures that you will obtain the same results as in this tutorial):

library(nosoi)  

#pExit
p_Exit_fct  <- function(t){return(0.08)}

#nContact
n_contact_fct = function(t){abs(round(rnorm(1, 0.5, 1), 0))}

#pTrans
p_Trans_fct <- function(t,p_max,t_incub){
    if(t < t_incub){p=0}
    if(t >= t_incub){p=p_max}
    return(p)
}
 
t_incub_fct <- function(x){rnorm(x,mean = 7,sd=1)}
p_max_fct <- function(x){rbeta(x,shape1 = 5,shape2=2)}

param_pTrans = list(p_max=p_max_fct,t_incub=t_incub_fct)

# Starting the simulation ------------------------------------

set.seed(805)
SimulationSingle <- nosoiSim(type="single", popStructure="none",
                             length.sim=100, max.infected=100, init.individuals=1, 
                             nContact=n_contact_fct,
                             param.nContact=NA,
                             timeDep.nContact=FALSE,
                             pExit = p_Exit_fct,
                             param.pExit=NA,
                             timeDep.pExit=FALSE,
                             pTrans = p_Trans_fct,
                             param.pTrans = param_pTrans,
                             timeDep.pTrans=FALSE,
                             prefix.host="H",
                             print.progress=FALSE)
#> Starting the simulation
#> Initializing ... running ...
#> done. 
#> The simulation has run for 40 units of time and a total of 111 hosts have been infected.

Once the simulation has finished, it reports the number of time units for which the simulation has run (40), and the maximum number of infected hosts (111). Note that the simulation has stopped here before reaching length.sim as it has crossed the max.infected threshold set at 100.

pExit.B

For pExit.B, we choose a value that depends on the “absolute” time of the simulation, for example cyclic climatic conditions (temperature). In that case, the function’s arguments should be t and prestime (the “absolute” time of the simulation), in that order:

p_Exit_fctB  <- function(t,prestime){(sin(prestime/(2*pi*10))+1)/16} #for a periodic function

The values of pExit.B across the “absolute time” of the simulation will be the following:

Since pExit.B is dependent on the simulation’s absolute time, do not forget to set timeDep.pExit.B to TRUE. Since there are no individual-based parameters, param.pExit.B=NA.

nContact.B

For nContact.B, we choose a constant function that will sample a value out of a provided range of possible values, each with a certain probability:

n_contact_fct.B = function(t){sample(c(0,1,2),1,prob=c(0.6,0.3,0.1))}

The distribution of nContact.B looks as follows:

At each time and for each infected host, nContact.B will be drawn anew. Remember that nContact.B, like the other core functions has to be function of t, even if t is not used. Since nContact.B is constant here, there is no use for the “absolute” time of the simulation nor for the individual-based parameters. So param.nContact.B=NA, and timeDep.nContact.B=FALSE.

pTrans.B

We choose pTrans.B in the form of a Gaussian function. It will reach its maximum value at a certain time point (mean) after initial infection and will subsequently decrease until it reaches 0:

p_Trans_fct.B <- function(t, max.time){
  dnorm(t, mean=max.time, sd=2)*5
}

Because each host is different (slightly different biotic and abiotic factors), you can expect each host to exhibit differences in the dynamics of infection, and hence the probability of transmission over time. Thus, max.time will be sampled for each host individually according to a certain distribution. max.time will be sampled from a normal distribution of parameters mean = 5 and sd = 1:

max.time_fct <- function(x){rnorm(x,mean = 5,sd=1)}

Note again that here max.time is a function of x and not t (not a core function but individual-based parameters), and x enters the function as the number of draws to make.

Taken together, the profile for pTrans for a subset of 200 individuals in the population will look as follows:

Since pTrans.B is not dependent on the “absolute” time of the simulation, timeDep.pTrans.B=FALSE. However, since we make use of individual-based parameters, we have to provide a param.pTrans as a list of functions. The name of each element of the list should have the same name as the core function (here pTrans.B) uses as argument, as shown here:

max.time_fct <- function(x){rnorm(x,mean = 5,sd=1)}

param_pTrans.B = list(max.time=max.time_fct)

Running

Once nosoiSim is set up, you can run the simulation (here the “seed” ensures that you will obtain the same results as in this tutorial):

library(nosoi)

#HostA ------------------------------------

#pExit
p_Exit_fct.A  <- function(t){return(0.08)}

#nContact
n_contact_fct.A = function(t){abs(round(rnorm(1, 0.5, 1), 0))}

#pTrans
p_Trans_fct.A <- function(t,p_max,t_incub){
    if(t < t_incub){p=0}
    if(t >= t_incub){p=p_max}
    return(p)
}
 
t_incub_fct <- function(x){rnorm(x,mean = 7,sd=1)}
p_max_fct <- function(x){rbeta(x,shape1 = 5,shape2=2)}

param_pTrans.A = list(p_max=p_max_fct,t_incub=t_incub_fct)

#Host B ------------------------------------

#pExit
p_Exit_fct.B  <- function(t,prestime){(sin(prestime/(2*pi*10))+1)/16}

#nContact
n_contact_fct.B = function(t){sample(c(0,1,2),1,prob=c(0.6,0.3,0.1))}

#pTrans
p_Trans_fct.B <- function(t, max.time){
  dnorm(t, mean=max.time, sd=2)*5
}
 
max.time_fct <- function(x){rnorm(x,mean = 5,sd=1)}

param_pTrans.B = list(max.time=max.time_fct)

# Starting the simulation ------------------------------------

set.seed(606)
SimulationDual <- nosoiSim(type="dual", popStructure="none",
                           length.sim=100, 
                           max.infected.A=100, 
                           max.infected.B=100, 
                           
                           init.individuals.A=1, 
                           init.individuals.B=0, 
                           
                           nContact.A=n_contact_fct.A,
                           param.nContact.A=NA,
                           timeDep.nContact.A=FALSE,
                           pExit.A=p_Exit_fct.A,
                           param.pExit.A=NA,
                           timeDep.pExit.A=FALSE,
                           pTrans.A=p_Trans_fct.A,
                           param.pTrans.A=param_pTrans.A,
                           timeDep.pTrans.A=FALSE,
                           prefix.host.A="H",
                           
                           nContact.B=n_contact_fct.B,
                           param.nContact.B=NA,
                           timeDep.nContact.B=FALSE,
                           pExit.B=p_Exit_fct.B,
                           param.pExit.B=NA,
                           timeDep.pExit.B=TRUE,
                           pTrans.B=p_Trans_fct.B,
                           param.pTrans.B=param_pTrans.B,
                           timeDep.pTrans.B=FALSE,
                           prefix.host.B="V",
                           
                           print.progress=FALSE)
#> Starting the simulation
#> Initializing ... running ...
#> done. 
#> The simulation has run for 43 units of time and a total of 101 (A) and 92 (B) hosts have been infected.

Once the simulation has finished, it reports the number of time units for which the simulation has run (43), and the maximum number of infected hosts A (101) and hosts B (92). Note that the simulation has stopped here before reaching length.sim as it has crossed the max.infected.A threshold set at 100.