midasdashboard: MiDAS dashboard shiny app

#####################################################
# For this one you need at least 20GB RAM AVAILABLE #
#####################################################

source("data-raw/helper-functions.R")
require("data.table")

data("mfg_functions")
metadata <- fread("data-raw/amplicon_data/biobank/metadata.txt", encoding = "Latin-1")
#this is hopefully FALSE, otherwise manually check which samples to remove:
any(duplicated(metadata$Sample))

#filter a few useless samples
metadata <- metadata[!Sample %chin% paste0("MQ201110-", 309:311)]

#### fix Plant and ID columns ####
metadata[grepl("extneg", tolower(LibID)), ID := "EXTNEG"]
metadata[grepl("extneg", tolower(LibID)), Plant := "CTRL"]
metadata[grepl("pcrpor", tolower(LibID)), ID := "PCRPOS"]
metadata[grepl("pcrpos", tolower(LibID)), Plant := "CTRL"]
metadata[grepl("pcrneg", tolower(LibID)), ID := "PCRNEG"]
metadata[grepl("pcrneg", tolower(LibID)), Plant := "CTRL"]
metadata <- metadata[!is.na(Plant)] #this removes the weird LibID samples: MQ181023-148, MQ181203-218, MQ181214-138, MQ190919-270
metadata <- metadata[!Sample %chin% "MQ201110-248"]
metadata[ID == "Lynetten", Plant := "Lynetten"]
metadata[ID == "Avedøre", Plant := "Avedøre"]
metadata[grepl("^Marselisborg", Plant), Plant := "Marselisborg"]

metadata[is.na(Line) | Line == "" | Line == "NA", Line := NA]
metadata[!Plant %chin% c("Damhusåen", "Marselisborg"), Line := NA]
metadata[grepl("^LIB-AAW-HC-U", LibID), Line := "U"]
metadata[grepl("^LIB-AAW-HC-O", LibID), Line := "O"]
metadata[, Plant := ifelse(!is.na(Line), paste0(Plant, "-", Line), Plant)]
#metadata[is.na(Plant) & is.na(Line), Plant := ID]

biobank_as <- amp_load(
  otutable = "data-raw/amplicon_data/biobank/ASVtable_midas481.zip",
  metadata = metadata
)

biobank_as <- amp_subset_samples(
  biobank_as,
  !grepl("ext|pcr|neg|pos", tolower(LibID)) & !Plant %in% "PCRNEG",
  normalise = TRUE
)

#this step uses an awful lot of memory
biobank_as <- filter_otus(biobank_as, 0.1)

biobank_as$metadata$Date <- lubridate::ymd(biobank_as$metadata$Date)
biobank_as$metadata <- fix_metadata(biobank_as$metadata)
biobank_as <- genusfunctions(
  biobank_as,
  function_data = mfg_functions
)

usethis::use_data(biobank_as, overwrite = TRUE)

#biobank_as_PeriodAvg <- periodAvg(biobank_as$metadata)
#usethis::use_data(biobank_as_PeriodAvg, overwrite = TRUE)