R/plot-spagedi.R

In lukembrowne/rSpagedi: Analysis of fine-scale spatial genetic structure data.

### Make a spatial autocorrelation plot with spagedi output

plotAutoCor <- function(spagediList, overlay = FALSE, color = "black",
                        max_dist){

  symbols <- 19
  
  perm <- spagediList$perm
  dist <- spagediList$dist
  dist <- as.numeric(spagediList$dist["Mean distance", ])
  
  if(!is.null(perm)){
    perm <- perm[, c(2:(ncol(perm) - 4))] # chop off last 3 columns and 1st col
    
    obs <- as.numeric(perm["Obs val", ])
    conf_hi <- as.numeric(perm["95%CI-sup", ])
    conf_low <- as.numeric(perm["95%CI-inf", ])
    
      # Closed symbol if permutation was significant
    sig <- apply(perm[c(8,9,10), ], 2, FUN = function(x) {any(x < 0.05)})
    symbols <- rep(1, ncol(perm))
    symbols[sig] <- 19
  } else{
    kin <- spagediList$kin
    kin <- kin[, c(2:(ncol(kin) - 4))]
    obs <- kin["ALL LOCI", ]
    conf_hi = 0
    conf_low = 0
  }
  
  if(overlay){
    par(new = TRUE)
    points(dist, obs, type = "b", pch = symbols, xlab = "", ylab = "",
           yaxt = "n", xaxt = "n", col = color, lwd = 2, lty = 1)
    if(!is.null(perm)){
      lines(dist, conf_hi, lty = 4, col = color)
      lines(dist, conf_low, lty = 4, col = color)
    }
    
  } else{
  
  plot(dist, obs, type = "b", pch = symbols, las = 1, 
       ylab = "Kinship", xlab = "Distance (m)", lwd = 2,
       ylim = c(min(obs, conf_low, na.rm = TRUE) * 1.1, 
                max(obs, conf_hi, na.rm = TRUE) * 1.1),
       xlim = c(0, max_dist), col = color)
  abline(h = 0, lty = 1, col = "grey50")
    if(!is.null(perm)){
      lines(dist, conf_hi, lty = 4, col = color)
      lines(dist, conf_low, lty = 4, col = color)
    }
  }
  
}