R/dibbler.R

In thibautjombart/dibbler: Investigation of Food-Borne Disease Outbreaks

## #' Modelling food-borne outbreaks using genetic data
## #'
## #' The function \code{dibbler} is used for inferring the transmission of food-borne diseases across a food distribution network using pathogen genome sequences sampled for a set of cases.
## #'
## #' !!! This package is still under development. Do not use it without contacting the author. !!!
## #'
## #'
## #' @author Thibaut Jombart \email{thibautjombart@@gmail.com}
## #'
## #' @export
## #' @importFrom igraph dfs
## #'
## #' @param x a list of the class 'dibbler.data' as returned by \code{\link{dibbler.data}}.
## #' @param graph.opt a list of options are returned by  \code{\link{dibbler.graph.opt}}.
## ## #' @param ... further arguments to be passed to other functions
## #'
## #' @seealso
## #' \describe{
## #' \item{\code{\link{dibbler.data}}}{to prepare the input data.}
## #' }
## #'
## #' @examples
## #'
## #' if(require(igraph)){
## #' ## generate graph from edge list
## #' Salmonella
## #' g <- graph.data.frame(Salmonella$graph)
## #'
## #' ## run dibbler
## #' out <- dibbler(dibbler.data(graph=g, group=Salmonella$cluster))
## #'
## #' ## check output
## #' names(out)
## #' head(out$freq)
## #' head(out$conf)
## #' out$graph
## #'
## #' ## plot results
## #' plot(out$graph, vertex.label="",
## #' main="dibbler: inferred pathogen distribution")
## #' }
## dibbler <- function(x=dibbler.data(), graph.opt=dibbler.graph.opt()){
##     ## CHECKS ##
##     check.data(x)

##     ## PERFORM  ##
##     freq <- list()
##     conf <- numeric()

##     ## for all internal nodes...
##     for(i in seq_along(x$lab.graph)){
##         ## get tree from the node
##         tree <- igraph::dfs(graph=x$graph, root=i,
##                     neimode="out", unreachable=FALSE)$order

##         ## remove NAs, keep only labels
##         tree <- names(tree)[!is.na(tree)]

##         ## identify tips
##         tips <- intersect(tree, x$lab.match)

##         ## get group frequencies
##         freq[[i]] <- table(x$group[tips])/length(tips)

##         ## get confidence measure
##         ## (prop of terminal nodes in tree)
##         conf[i] <- length(tips)/length(tree)
##     }


##     ## SET OUTPUT GRAPH ATTRIBUTES ##
##     graph <- set.graph.opt(x$graph, graph.opt, freq=freq, conf=conf)

##     ## SHAPE/RETURN OUTPUT ##
##     names(freq) <- names(conf) <- x$lab.graph
##     out <- list(freq=freq, conf=conf, graph=graph)

##     return(out)
## } # end dibbler