heatmap error in Galaxy tool

[bernt-matthias]

I'm currently debugging the ampvis2 heatmap Galaxy tool (still based on 2.7.22) which gives an error in the CI since two weeks ago:

Error in if (y[["In situ"]] == "Positive" | (any(y[["In situ"]] %in% c("Variable",  : 
  argument is of length zero
Calls: amp_heatmap -> extractFunctions -> lapply -> FUN -> lapply -> FUN

traceback() gives me

6: FUN(X[[i]], ...)
5: lapply(c(x[["properties"]], x[["metabolism"]]), function(y) {
       if (y[["In situ"]] == "Positive" | (any(y[["In situ"]] %in% 
           c("Variable", "Not Assessed")) & y[["Other"]] == "Positive")) {
           return("POS")
       }
       if (y[["In situ"]] == "Negative" | (y[["In situ"]] == "Not Assessed" & 
           y[["Other"]] == "Negative")) {
           return("NEG")
       }
       if (y[["In situ"]] == "Variable" | (y[["In situ"]] == "Not Assessed" & 
           y[["Other"]] == "Variable")) {
           return("VAR")
       }
       if (all(c(y[["In situ"]], y[["Other"]]) %in% "Not Assessed")) {
           return("NT")
       }
   })
4: FUN(X[[i]], ...)
3: lapply(FGList, function(x) {
       outList <- lapply(c(x[["properties"]], x[["metabolism"]]), 
           function(y) {
               if (y[["In situ"]] == "Positive" | (any(y[["In situ"]] %in% 
                   c("Variable", "Not Assessed")) & y[["Other"]] == 
                   "Positive")) {
                   return("POS")
               }
               if (y[["In situ"]] == "Negative" | (y[["In situ"]] == 
                   "Not Assessed" & y[["Other"]] == "Negative")) {
                   return("NEG")
               }
               if (y[["In situ"]] == "Variable" | (y[["In situ"]] == 
                   "Not Assessed" & y[["Other"]] == "Variable")) {
                   return("VAR")
               }
               if (all(c(y[["In situ"]], y[["Other"]]) %in% "Not Assessed")) {
                   return("NT")
               }
           })
       c(Genus = gsub("^Ca ", "Ca_", x[["name"]]), MiDAS = "POS", 
           outList)
   })
2: extractFunctions(function_data)
1: amp_heatmap(d, normalise = TRUE, tax_aggregate = "Genus", tax_add = NULL, 
       tax_show = 10, showRemainingTaxa = FALSE, tax_empty = "best", 
       plot_values = TRUE, plot_values_size = 4, plot_colorscale = "log10", 
       plot_na = FALSE, measure = "mean", min_abundance = 0.1, textmap = FALSE, 
       plot_functions = TRUE, functions = c("Filamentous"), rel_widths = c(0.75, 
           0.25))

Call is:

plot <- amp_heatmap(d, normalise = TRUE, tax_aggregate = "Genus", tax_add = NULL, tax_show = 10, showRemainingTaxa = FALSE, tax_empty = "best", plot_values = TRUE, plot_values_size = 4, plot_colorscale = "log10", plot_na = FALSE, measure = "mean", min_abundance = 0.1, textmap = FALSE, plot_functions = TRUE, functions = c("MiDAS", "Filamentous", "AOB", "NOB", "GAO"), rel_widths = c(0.75, 0.25))

[KasperSkytte]

Hi Matthias

Thanks for notifying. I've debugged it and it seems there was simply a typo in the MiDAS field guide database from where amp_heatmap pulls the functional information (one entry, Sellimonas). It's fixed in the database, working fine here. In the future, if getting errors related to functional information you can always just use the included data set with ampvis2 midasfunctions and passing it along with the function_data argument. It's not updated often though, the website is the most up-to-date.

[bernt-matthias]

Thanks @KasperSkytte this also fixed the CI problem