Update import of bugphyzz and how to make signatures

NAMESPACE

@@ -5,6 +5,7 @@ export(fattyAcidComposition)
 export(getBugAnnotations)
 export(getBugphyzzSignatures)
 export(importBugphyzz)
+export(makeSignatures)
 export(physiologies)
 export(showPhys)
 export(whichAttr)

R/bugphyzz.R

@@ -14,39 +14,154 @@
 #' @examples
 #'
 #' bp <- importBugphyzz()
-#'
-#' ## Check available groups of attributes
-#' unique(bp$Attribute_group)
-#'
-#' ## Filter only for growth temperature
-#' gt <- bp[which(bp$Attribute_group == 'growth temperature'), ]
-#'
-#' ## Create signatures with taxids at the species level
-#' gt_sigs <- getBugphyzzSignatures(gt, tax.id.type = 'NCBI_ID', tax.level = 'species')]
-#' lapply(gt_sigs, function(x) head(x))
-#'
+#' names(bp)
 #'
 importBugphyzz <- function(version = 'devel', force_download = FALSE) {
-  if (version == 'devel')
-    url <- 'https://github.com/waldronlab/bugphyzzExports/raw/sdgamboa/update-workflow/bugphyzz_export.tsv'
-  rpath <- .getResource(
-    rname = 'bugphyzz_export.tsv', url = url, verbose = TRUE,
-    force = force_download
+  types <- c("multistate", "binary", "numeric")
+  urls <- paste0(
+    "https://github.com/waldronlab/bugphyzzExports/raw/sdgamboa/phylo/bugphyzz_",
+    types,
+    ".csv"
   )
-  thr <- .thresholds()
-  dat <- utils::read.table(rpath, header = TRUE, sep = '\t') |>
-    dplyr::mutate(Score = round(.data$Score, digits = 3)) |>
-    dplyr::mutate(Frequency = dplyr::case_when(
-      .data$Score == 1 ~ 'always',
-      .data$Score >= 0.9 & .data$Score < 1 ~ 'usually',
-      .data$Score >= 0.5 & .data$Score < 0.9 ~ 'sometimes',
-      .data$Score > 0 & .data$Score < 0.5 ~ 'rarely',
-      .data$Score == 0  ~ 'never'
-    )) |>
-    dplyr::mutate(
-      Attribute_source = ifelse(.data$Evidence == 'inh', NA, .data$Attribute_source)
+  names(urls) <- types
+  if (version == 'devel') {
+    output <- vector("list", length(urls))
+    for (i in seq_along(output)) {
+      message("Importing ", names(urls)[i], " data...")
+      names(output)[i] <- names(urls)[i]
+      rpath <- .getResource(
+        rname = paste0("bugphyzz_", names(urls)[i], ".tsv"),
+        url = urls[i], verbose = TRUE, force = force_download
+      )
+      output[[i]] <- utils::read.csv(rpath, header = TRUE, skip = 1)
+    }
+  }
+
+  output <- lapply(output, function(x) split(x, x$Attribute_group))
+  output <- purrr::list_flatten(output)
+  names(output) <- purrr::map_chr(output, ~ unique(.x$Attribute_group))
+  return(output)
+}
+
+
+#' Make bugs signatures
+#'
+#' \code{makeSignatures} Creates signatures for a list of bugphyzz
+#' data.frames imported with \code{importBugphyzz}
+#'
+#' @param dat A data.frame.
+#' @param tax_id_type A character string. Valid options: NCBI_ID, Taxon_name.
+#' @param tax_level A character vector. Taxonomic rank. Valid options:
+#' kingdom, phylum, class, order, family, genus, species, strain.
+#' They can be combined. "mixed" is equivalent to select all valid ranks.
+#' @param evidence A character vector. Valid options: exp, igc, nas, tas, tax, asr.
+#' They can be combined. Default is all.
+#' @param frequency A character vector. Valid options: always, usually,
+#' sometimes, rarely, unknown. They can be combiend. Default value is all but
+#' rarely.
+#' @param min_size Minimun number of bugs in a signature. Default is 10.
+#' @param min Minimum value inclusive. Only for numeric attributes. Default is NULL.
+#' @param max Maximum value inclusive. Only for numeric attributes. Default is NULL.
+#'
+#' @return A list of character vector with the IDs of the bugs.
+#' @export
+#'
+#' @examples
+#'
+#' bp <- importBugphyzz()
+#' sigs <- lapply(bp, makeSignatures)
+#' sigs <- purrr::list_flatten(sigs)
+#'
+makeSignatures <- function(
+    dat, tax_id_type = "NCBI_ID",
+    tax_level = "mixed",
+    evidence = c("exp", "igc", "tas", "nas", "tax", "asr"),
+    frequency = c("always", "usually", "sometimes", "unknown"),
+    min_size = 10, min = NULL, max = NULL
+) {
+  attr_type <- unique(dat$Attribute_type)
+  if (tax_level == "mixed") {
+    tax_level <- c(
+      "kingdom", "phylum", "class", "order", "family", "genus", "species",
+      "strain"
     )
-  dplyr::left_join(dat, thr, by = c('Attribute_group', 'Attribute'))
+  }
+  dat <- dat |>
+    dplyr::filter(Rank %in% tax_level) |>
+    dplyr::filter(.data$Evidence %in% evidence) |>
+    dplyr::filter(.data$Frequency %in% frequency)
+  if (!nrow(dat)) {
+    warning(
+      "Not enough data for creating signatures. Try different filtering options",
+      call. = FALSE
+    )
+    return(NULL)
+  }
+  if (attr_type %in% c("multistate-intersection", "binary")) {
+    s <- .makeSignaturesDiscrete(dat = dat, tax_id_type = tax_id_type)
+  } else if (attr_type %in% c("range", "numeric")) {
+    s <- .makeSignaturesNumeric(
+      dat = dat, tax_id_type = tax_id_type, min = min, max = max
+    )
+  }
+  output <- purrr::keep(s, ~ length(.x) >= min_size)
+  if (!length(output)) {
+    warning(
+      "Not enough data for creating signatures. Try different filtering options",
+      call. = FALSE
+    )
+  }
+  return(output)
+}
+
+.makeSignaturesDiscrete <- function(dat, tax_id_type = "NCBI_ID") {
+  dat |>
+    dplyr::mutate(
+      Attribute = paste0("bugphyzz:", .data$Attribute, "|", .data$Attribute_value)
+    ) |>
+    {\(y) split(y, y$Attribute)}() |>
+    lapply(function(x) unique(x[[tax_id_type]]))
+}
+
+.makeSignaturesNumeric <- function(
+    dat, tax_id_type = "NCBI_ID", min = NULL, max = NULL
+) {
+  if (!is.null(min) || !is.null(max)) {
+    if (is.null(min)) {
+      message("Minimum unespecified. Using ", min(dat$Attribute_value), ".")
+      min <- min(dat$Attribute_value)
+    }
+    if (is.null(max)) {
+      message("Maximum unespecified. Using ", max(dat$Attribute_value), ".")
+      max <- max(dat$Attribute_value)
+    }
+    dat <- dat |>
+      dplyr::filter(
+        .data$Attribute_value >= min & .data$Attribute_value <= max
+      ) |>
+      dplyr::mutate(
+        Attribute = paste0("bugphyzz:", .data$Attribute, "| >=", min, " & <=", max)
+      )
+  } else {
+    thr <- .thresholds() |>
+      dplyr::filter(.data$Attribute_group == unique(dat$Attribute_group))
+    attr_name <- thr$Attribute
+    min_values <- thr$lower
+    max_values <- thr$upper
+    dat$tmp_col <- NA
+    for (i in seq_along(attr_name)) {
+      if (is.na(min_values[i]))
+        min_values[i] <- min(dat$Attribute_value) - 0.01
+      if (is.na(max_values[i]))
+        max_values[i] <- max(dat$Attribute_value)
+      pos <- which(dat$Attribute_value > min_values[i] & dat$Attribute_value <= max_values[i])
+      dat$tmp_col[pos] <- attr_name[i]
+      dat$Attribute[pos] <- paste0("bugphyzz:", dat$Attribute[pos], "|", attr_name[i], "| > ", round(min_values[i], 2), " & <= ", max_values[i])
+    }
+  }
+  dat |>
+    {\(y) split(y, y$Attribute)}() |>
+    lapply(function(x) unique(x[[tax_id_type]]))
 }
 
 #' Get bugphyzz signatures
@@ -235,7 +350,7 @@ whichAttrGrp <- function(bp) {
       unit = ifelse(is.na(.data$unit), '', .data$unit)
     ) |>
     dplyr::mutate(Attribute_range = paste0(range, unit)) |>
-    dplyr::select(
+    dplyr::relocate(
       .data$Attribute_group, .data$Attribute, .data$Attribute_range
     )
 }

inst/extdata/thresholds.tsv

@@ -1,32 +1,32 @@
 Attribute_group	Attribute	lower	upper	unit
 coding genes	very small	NA	473	NA
-coding genes	small	474	600	NA
-coding genes	average	601	6000	NA
-coding genes	very large	60001	NA	NA
+coding genes	small	473	600	NA
+coding genes	average	600	6000	NA
+coding genes	very large	6000	NA	NA
 genome size	small	NA	490885	bp
-genome size	average	490886	998123	bp
-genome size	large	998124	6997434	bp
-genome size	very large	6997435	NA	bp
-growth temperature	psychrophile	NA	24.9	C
+genome size	average	490885	998123	bp
+genome size	large	998123	6997434	bp
+genome size	very large	6997434	NA	bp
+growth temperature	psychrophile	NA	25	C
 growth temperature	mesophile	25	45	C
-growth temperature	thermophile	46	60	C
-growth temperature	hyperthermophile	61	NA	C
+growth temperature	thermophile	45	60	C
+growth temperature	hyperthermophile	60	NA	C
 length	small	NA	3.8	μm
-length	average	3.9	22	μm
-length	large	23	60	μm
-length	very large	61	NA	μm
+length	average	3.8	22	μm
+length	large	22	60	μm
+length	very large	60	NA	μm
 mutation rate per site per generation	slow	NA	2.92	NA
-mutation rate per site per generation	medium	2.93	16	NA
-mutation rate per site per generation	fast	17	NA	NA
+mutation rate per site per generation	medium	2.92	16	NA
+mutation rate per site per generation	fast	16	NA	NA
 mutation rate per site per year	slow	NA	7.5	NA
-mutation rate per site per year	medium	7.6	20	NA
-mutation rate per site per year	medium fast	21	54.2	NA
-mutation rate per site per year	fast	54.3	NA	NA
+mutation rate per site per year	medium	7.5	20	NA
+mutation rate per site per year	medium fast	20	54.2	NA
+mutation rate per site per year	fast	54.2	NA	NA
 optimal ph	acidic	NA	6	NA
 optimal ph	neutral	6	8	NA
 optimal ph	alkaline	8	9.76	NA
 optimal ph	very alkaline	9.76	NA	NA
 width	small	NA	0.9	μm
-width	average	0.91	3.5	μm
-width	large	3.51	12	μm
-width	very large	13	NA	μm
+width	average	0.9	3.5	μm
+width	large	3.5	12	μm
+width	very large	12	NA	μm