Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add lefesrPlotClad function #63

Merged
merged 12 commits into from
Sep 30, 2024
3 changes: 3 additions & 0 deletions NAMESPACE
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,9 @@
export(get_terminal_nodes)
export(lefsePlotFeat)
export(lefser)
export(lefserAllRanks)
export(lefserPlot)
export(lefserPlotClad)
export(relativeAb)
import(SummarizedExperiment)
import(ggplot2)
Expand All @@ -14,6 +16,7 @@ importFrom(coin,wilcox_test)
importFrom(dplyr,"%>%")
importFrom(dplyr,arrange)
importFrom(dplyr,mutate)
importFrom(ggtree,"%<+%")
importFrom(methods,as)
importFrom(methods,is)
importFrom(stats,kruskal.test)
Expand Down
6 changes: 6 additions & 0 deletions R/lefser.R
Original file line number Diff line number Diff line change
Expand Up @@ -331,6 +331,12 @@ lefser <-
attr(res_scores, "blk") <- blockCol
attr(res_scores, "method") <- method
attr(res_scores, "lgroupf") <- lgroupf[1]
attr(res_scores, "case") <- lgroupf[2]

## Some more attributes to create the cladogram.
pathStrings <- .selectPathStrings(relab, res_scores)
attr(res_scores, "pathStrings") <- pathStrings
attr(res_scores, "tree") <- .toTree(pathStrings)
res_scores
}

Expand Down
326 changes: 326 additions & 0 deletions R/lefserPlotClad.R
Original file line number Diff line number Diff line change
@@ -0,0 +1,326 @@

# Functions for plotting a cladogram --------------------------------------

#' LEfSer plot cladogram
#'
#' \code{lefserPlotClad} plots a cladogram from the results of
#' `lefser` or `lefserAllRanks`
#'
#' @param df An object of class "lefser_df" or "lefesr_df_all".
#' @param colors Colors corresponding to class 0 and 1.
#' Options: "c" (colorblind), "l" (lefse), "g" (greyscale).
#' Defaults to "c". This argument also accepts a character(2) with two color names.
#' @param showTipLabels Logical. If TRUE, show tip labels. Default is FALSE.
#' @param showNodeLabels Options: "p" = phylum, "c" = class, "o" = order,
#' "f" = family, "g" = genus. It can accept several options, e.g.,
#' c("p", "c").
#'
#' @importFrom ggtree %<+%
#'
#' @return A ggtree object.
#' @export
#'
#' @examples
#' data("zeller14")
#' z14 <- zeller14[, zeller14$study_condition != "adenoma"]
#' tn <- get_terminal_nodes(rownames(z14))
#' z14tn <- z14[tn, ]
#' z14tn_ra <- relativeAb(z14tn)
#' resAll <- lefserAllRanks(relab = z14tn_ra, groupCol = "study_condition")
#' ggt <- lefserPlotClad(df = resAll)
lefserPlotClad <- function(
df, colors = "c", showTipLabels = FALSE, showNodeLabels = "p"
) {
inputClass <- class(df)[1]
if (inputClass == "lefser_df") {
message("Woriking with lefser_df. Consider using lefserAll.")
# df$features <- .extracTips(df$features)
} else if (inputClass == "lefser_df_all") {
message("Working with lefser_df_all")
## .extractTips should be use here as well
## The feature names format should use full taxonomy
} else {
stop(
"You need an object of class 'lefser_df_all'",
call. = FALSE
)
}

df$features <- .extracTips(df$features)

colors <- .selectPalette(colors)
tree <- attr(df, "tree")
controlVar <- attr(df, "lgroupf")
caseVar <- attr(df, "case")

res <- df |>
dplyr::mutate(
sample = dplyr::case_when(
## This assumes positive values always mean enriched in
## the case condition.
.data[["scores"]] > 0 ~ .env[["caseVar"]],
TRUE ~ .env[["controlVar"]]
)
) |>
dplyr::mutate(abs = abs(.data[["scores"]])) |>
as.data.frame()

labels <- c(tree$tip.label, tree$node.label)
res$node <- match(res$features, labels)
dat <- dplyr::relocate(res, node)

internalNodes <- ape::Ntip(tree) + 1:ape::Nnode(tree)
collapseThem <- purrr::map_int(internalNodes, ~ {
chNods <- treeio::offspring(.data = tree, .node = .x, type = "tips")
if (any(chNods %in% dat$node)) {
return(NA)
} else {
return(.x)
}
}) |>
purrr::discard(is.na)

nodLab <- match.arg(
arg = showNodeLabels,
choices = c("p", "c", "o", "f", "g"),
several.ok = TRUE
)
nodLabRgx <- paste0("[", paste0(nodLab, collapse = ""), "]__")
treeData <- dat |>
dplyr::mutate(
showNodeLabs = dplyr::case_when(
grepl(nodLabRgx, features) ~ features,
TRUE ~ NA
)
)
# return(treeData)

gt <- ggtree::ggtree(
tree, layout = "circular", branch.length = "none", size = 0.2
) %<+% treeData

if (showTipLabels) {
gt <- gt +
ggtree::geom_tiplab(
mapping = ggtree::aes(label = features), size = 2,
geom = "text", na.rm=TRUE
)
}

gt2 <- gt +
ggtree::geom_tippoint(
mapping = ggtree::aes(fill = sample, size = abs), shape = 21,
na.rm=TRUE
) +
ggtree::geom_nodepoint(
mapping = ggtree::aes(fill = sample, size = abs), shape = 21,
na.rm = TRUE
) +
ggrepel::geom_label_repel(
mapping = ggtree::aes(label = showNodeLabs),
na.rm = TRUE
) +
ggtree::scale_fill_manual(
values = colors, breaks = c(controlVar, caseVar),
name = "Sample", na.value = NA
) +
ggplot2::scale_size(name = "Absolute\nscore") +
ggtree::theme(legend.position = "right")
for (i in collapseThem) {
gt2 <- withCallingHandlers(
ggtree::collapse(gt2, node = i),
warning = function(w) {
if (grepl("collapse", w$message)) {
invokeRestart("muffleWarning")
}
}
)
}
return(gt2)
}

# Run lefser at all taxonomic levels --------------------------------------

#' Run lefser on all taxonomic levels
#'
#' @param relab A SummarizedExperiment.
#' @param ... Arguments passed to the \code{lefser} function.
#'
#' @return An object of class 'lefser_df_all' and 'data.frame'.
#' @export
#'
#' @examples
#'
#' data("zeller14")
#' z14 <- zeller14[, zeller14$study_condition != "adenoma"]
#' tn <- get_terminal_nodes(rownames(z14))
#' z14tn <- z14[tn, ]
#' z14tn_ra <- relativeAb(z14tn)
#'
#' resAll <- lefserAllRanks(relab = z14tn_ra, groupCol = "study_condition")
#'
lefserAllRanks <- function(relab,...) {
## Feature names should have the full taxonomy
se <- .rowNames2RowData(relab)
seL <- mia::splitByRanks(se)
## The kingdom level is not needed
## The mia package doesn't support strain.
seL <- seL[names(seL) != "kingdom"]
seL <- purrr::map(seL, ~ {
seVar <- .x
rowDat <- as.data.frame(SummarizedExperiment::rowData(seVar))
rowDat <- purrr::discard(rowDat, function(x) all(is.na(x)))
rowDat <- S4Vectors::DataFrame(rowDat)
SummarizedExperiment::rowData(seVar) <- rowDat
seVar
})
for (i in seq_along(seL)) {
rownames(seL[[i]]) <- .lognRowNames(seL[[i]])
}

res <- seL |>
purrr::map(function(x, ...) lefser(relab = x,...), ...) |>
dplyr::bind_rows()
resOriginal <- lefser(relab, ...)
## Get only tip names (full names with full taxonomy are too long).
# resOriginal$features <- stringr::str_extract(
# resOriginal$features, "[^|]+$"
# )
res <- res |>
## Avoid repeating features.
dplyr::filter(!.data[["features"]] %in% resOriginal$features) |>
## Features not supported by mia are added (strain, OTUs, etc.)
dplyr::bind_rows(resOriginal)

controlVar <- attr(resOriginal, "lgroupf")
caseVar <- attr(resOriginal, "case")

class(res) <- c("lefser_df_all", class(res))

## These pathStrings could be used in the plotting function instead (or not)
pathStrings <- .selectPathStrings(relab, res)
attr(res, "pathStrings") <- pathStrings
attr(res, "tree") <- .toTree(pathStrings)

attr(res, "lgroupf") <- controlVar
attr(res, "case") <- caseVar
return(res)
}

## Add taxonomic information to rowData
## This step is necessary for mia to work
.rowNames2RowData <- function(x) {
sdgamboa marked this conversation as resolved.
Show resolved Hide resolved
se <- x
taxonomy <- .getTaxonomyFromPathStr(rownames(se))
dataFrame <- data.frame(tax = taxonomy) |>
tidyr::separate(
col = "tax", into = paste0("col", 1:10), # Number of taxa is usually seven, so 10 should be more than enough.
sep = "\\|", extra = "merge", fill = "right"
) |>
purrr::discard(~ all(is.na(.x)))
## purrr::map_chr ensures that the a single letter is used per column.
## Having two or more letters would trigger and error message from map_chr.
firstLetter <- purrr::map_chr(dataFrame, ~ {
taxLvl <- stringr::str_extract(.x, "\\w__")
unique(taxLvl[which(!is.na(taxLvl))])
})
rankNames <- dplyr::case_when(
firstLetter == "k__" ~ "kingdom",
firstLetter == "p__" ~ "phylum",
firstLetter == "c__" ~ "class",
firstLetter == "o__" ~ "order",
firstLetter == "f__" ~ "family",
firstLetter == "g__" ~ "genus",
firstLetter == "s__" ~ "species",
firstLetter == "t__" ~ "strain",
)
colnames(dataFrame) <- rankNames
DF <- S4Vectors::DataFrame(dataFrame)
SummarizedExperiment::rowData(se) <- DF
return(se)
}

## This functions makes sure that only the taxonomy
## is used for the rowData.
## OTU's or other non-typical taxonomic ranks will not be included.
.getTaxonomyFromPathStr <- function(pathStrings) {
rgx <- "^k__[^|]+\\|p__[^|]+\\|c__[^|]+\\|o__[^|]+\\|f__[^|]+(\\|g__[^|]+)?(\\|s__[^|]+)?(\\|t__[^|]+)?"
stringr::str_extract(pathStrings, pattern = rgx)
}

## This function selects pathStrings containing only
## taxa that is differentiallty abundant
.selectPathStrings <- function(se, res) {
pathStrings <- rownames(se)
index <- res$features |>
purrr::map(~ which(stringr::str_detect(pathStrings, .x))) |>
unlist() |>
unique() |>
sort()
pathStrings <- pathStrings[index]
return(pathStrings)
}

# Create cladogram --------------------------------------------------------
## Convert a character vector with pathStrings into a cladogram
## These could come from the rownames of a SummarizedExperiment with
## terminal nodes
.toTree <- function(pathStrs) {
edgeDF <- pathStrs |>
purrr::map(.pathString2EdgeList) |>
dplyr::bind_rows() |>
dplyr::distinct()
tipLabels <- stringr::str_extract(pathStrs, "[^|]+$")
nodeLabels <- unique(edgeDF$from)
idMap <- 1:(length(tipLabels) + length(nodeLabels))
names(idMap) <- c(tipLabels, nodeLabels)
edgeMat <- matrix(
data = c(idMap[edgeDF$from], idMap[edgeDF$to]),
ncol = 2
)
tr <- list(
edge = edgeMat,
tip.label = tipLabels,
node.label = nodeLabels,
Nnode = length(nodeLabels),
Ntip = length(tipLabels)
)
class(tr) <- "phylo"
tr
}

## Helper function for .toTree
## Input is a single path string, e.g., "k__bacteria|p_Fusobacteria..."
.pathString2EdgeList <- function(pathStr) {
pathStrRoot <- stringr::str_c("ROOT|", pathStr)
chr_vct <- stringr::str_split(pathStrRoot, "\\|")[[1]]
data.frame(
from = chr_vct[1:length(chr_vct)-1],
to = chr_vct[2:length(chr_vct)]
)
}

## This function extracts only the last element of the taxonomy
.extracTips <- function(pathStrs) {
stringr::str_extract(pathStrs, "[^|]+$")
}


# Utils -------------------------------------------------------------------
.lognRowNames <- function(se) {
dat <- SummarizedExperiment::rowData(se) |>
as.data.frame() |>
tibble::rownames_to_column(var = "rowname") |>
dplyr::relocate(.data[["rowname"]])
lastColLgl <- all(dat[[colnames(dat)[ncol(dat)]]] == dat[["rowname"]])
if (lastColLgl) {
dat <- dat[, which(colnames(dat) != "rowname")]
output <- dat |>
tidyr::unite(
col = "features", 1:tidyselect::last_col(),
sep = "|", remove = TRUE,
) |>
dplyr::pull(.data[["features"]])
}
return(output)
}
18 changes: 18 additions & 0 deletions inst/scripts/cladogramPlot.R
Original file line number Diff line number Diff line change
@@ -0,0 +1,18 @@
suppressPackageStartupMessages(library(lefser))
data("zeller14")
z14 <- zeller14[, zeller14$study_condition != "adenoma"]
tn <- get_terminal_nodes(rownames(z14))
z14tn <- z14[tn, ]
z14tn_ra <- relativeAb(z14tn)

resAll <- lefserAllRanks(relab = z14tn_ra, groupCol = "study_condition")
ggt <- lefserPlotClad(df = resAll)
# y
# z <- lefserPlotClad(df = resAll, showTipLabels = TRUE, showNodeLabels = c("c"))
# z
# sessioninfo::session_info()


# res <- lefser(z14tn_ra, groupCol = "study_condition")
# x <- lefserPlotClad(df = res)
# x
Loading