R/msavisr.R
In vragh/seqvisr: Biological Sequence Visualization and Auxiliary Functions in R
Defines functions
msavisr
Documented in
msavisr
#'Multiple sequence alignment (MSA) visualization and manipulation thereof.
#'
#'Given a fasta formatted MSA, msavisr will attempt to produce a visualization of the MSA
#'with matches, mismatches, gaps, and (optional) regions of interest (ROIs) being highlighted
#'in different colors. The matches, mismatches, and gaps are calculated by comparing all other
#'sequences to a designated reference sequence within the MSA. ROIs are specified manually (see
#'below).
#'
#'
#' @usage msavisr(mymsa = NULL, myref = NULL, mypath = NULL,
#' refontop = TRUE, myroi = NULL, hnon = NULL, hmat = NULL,
#' hroi = NULL, wnon = NULL, wmat = NULL, wroi = NULL,
#' anon = NULL, amat = NULL, aroi = NULL, basecolors = NULL,
#' roicolors = NULL, cbfcols = TRUE)
#'
#' @param mymsa (character string, mandatory) the name of the fasta-formatted file containing the
#' multiple sequence alignment (MSA). The full path to the file can also be provided here (in which
#' case, the mypath argument must be set to NULL).
#'
#' @param myref (character string, mandatory) the fasta header (not including the ">") of one of the
#' sequences in the MSA which act as the reference sequence.
#'
#' @param mypath (character string, optional) the path to the directory containing the MSA. (Default
#' : looks in the current directory.)
#'
#' @param refontop (boolean, optional) should the reference sequence be placed at the top of the MSA plot
#' (default) or at the bottom? (Set to FALSE for bottom.)
#'
#' @param myroi (list of vectors, optional) the user can manually specify regions of interest (ROI;
#' i.e., positions in the MSA) that they wish to highlight manually via myroi. This must be supplied
#' as a list of vectors wherein each vector is of the format c(seqname, pos, desc), or c(seqname, pos),
#' or c(pos, desc), or c(pos). The seqname (name of a specific sequence in the MSA) and the desc
#' (description of the feature) are optional. pos (i.e., the positions that denote the ROI) can be a
#' single integer (e.g., 100), or a range of integers (1:10). Each such ROI vector can also hold one
#' or more (and combinations) of single integer pos values and integer range pos values (e.g., something
#' like c("Seq1", 1:10, 100, "Helix")). Thus, for example, if a SNP at position 100 in "seq2" and a the
#' region of nucleotides/amino-acids 20:30 representing a domain in all sequences were to be ROIs, they
#' would be supplied like so list(c("seq2", 100, "SNP"), c(20:30, "domain")). ROIs can be assigned any
#' color by the user (see argument roicolors); colors are assigned automatically otherwise (optionally
#' from a color blind-friendly palette; see argument cbfcols).
#'
#' @param hnon (double, optional) the height of the feature "bar(s)" for all mismatches and gaps in the MSA. (Default: 0.4.)
#'
#' @param hmat (double, optional) the height of the feature "bar(s)" for all matches in the MSA. (Default: 0.4.)
#'
#' @param hroi (double, optional) the height of the feature "bar(s)" for all ROIs in the MSA. (Default: 0.4.)
#'
#' @param wnon (double, optional) the width of the feature "bar(s)" for all mismatches and gaps in the MSA. (Default: 2.0.)
#'
#' @param wmat (double, optional) the width of the feature "bar(s)" for all matches in the MSA. (Default: 1.0.)
#'
#' @param wroi (double, optional) the width of the feature "bar(s)" for all ROIs in the MSA. (Default: 4.0.)
#'
#' @param anon (double, optional) the transparency of the feature "bar(s)" for all mismatches and gaps in the MSA. (Default: 1.0.)
#'
#' @param amat (double, optional) the transparency of the feature "bar(s)" for all matches in the MSA. (Default: 1.0.)
#'
#' @param aroi (double, optional) the transparency of the feature "bar(s)" for all ROIs in the MSA. (Default: 1.0.)
#'
#' @param basecolors (vector of 3 character strings, optional) the colors for the matches, mismatches, and gaps can optionally be
#' supplied by the user via this argument. Defaults are c("gray", "black", "white") when no colors are supplied and cbfcols (see
#' below) is set to FALSE. If cbfcols == TRUE, and no colors are supplied by the user then a palette from viridis is chosen.
#'
#' @param roicolors (vector of n character strings, optional) user-specified colors for the ROI features can be supplied via this
#' argument. As many colors as there are ROIs need to be supplied, and the order of the colors should correspond to the order of the
#' ROIs in the input list. If too few colors are supplied, colors are reused; if too many are supplied, the last few colors will not
#' be used. Defaults are chosen automacially from grDevices::color() if cbfcols == FALSE, and from viridis otherwise.
#'
#' @param cbfcols (boolean, optional) allows for the user to choose whether the automatic coloring scheme used should be color-blind
#' friendly. (Default: TRUE; set to FALSE to use non-color blind-friendly colors.)
#'
#' @details
#' msavisr plots the matches as a separate geom_tile() layer, the gaps + mismatches as a geom_tile() layer, and the ROIs as a
#' separate geom_tile() layer (if ROIs are supplied).
#'
#' The user will have to tweak the values for the widths and heights of these layers (via the arguments outlined above) to achieve
#' the desired visualization "effects". In general, it is advisable to set the widths of the mismatches + gaps (and/or ROIs, if any;
#' so wnon and wroi respectively) larger than that of the matches (wmat). The heights can be increased if necessary. Altering the
#' transparency levels does not seem to be very useful. Note: altering the transparency levels does not update the transparency of
#' the colors shown in the legend!!
#'
#' ROIs are especially useful for visualizing features such as single nucleotide polymorphisms (SNPs) in nucleotide MSAs and other
#' such isolated features that might normally become "buried" within the bulk of the sequence. This can be easily achieved by
#' indicating the SNPs position as an ROI and jacking up its width (wroi) and/or height (hroi) values.
#'
#' @return A ggplot2 object is returned to the parent environment for plotting and/or further downstream processing/manipulation.
#'
#' @note
#' The only issue with specifying ROIs in the manner implemented here is that, if for instance, a ROI in "seq2" at position 100
#' and in "seq4" at the same position need to be highlighted, it cannot be supplied like so c("seq2", "seq4", 100), and instead must
#' be supplied as two separate vectors c("seq2", 100), c("seq4", 100). The lowdown: specifying common ROIs in a SUBSET of the MSA can
#' be a tedious process. Unfortunately, as of now, no internal workarounds have been implemented.
#'
#' @examples
#' \dontrun{
#' #Input data
#' testmsa <- system.file("extdata", "testaln_mrna.fasta", package = "seqvisr", mustWork = TRUE)
#'
#' #Basic visualization
#' msavisr(mymsa = testmsa, myref = "Ref0") #No ROIs
#'
#' #Defining an example ROI
#' testroi <- list(c("Ref0", 100:110, "Ref0 Domain1"), c(14, "Pseudouridine"),
#' c(20:30, "Domain2"), c("Seq2", 55, "SNP"))
#'
#' #MSA with ROIs
#' msavisr(mymsa = testmsa, myref = "Ref0", myroi = testroi)
#' }
#'
#' @importFrom grDevices colors
#'
#' @importFrom utils read.table
#'
#' @importFrom viridis viridis
#'
#' @importFrom tidyr separate_rows
#'
#' @importFrom magrittr %>% %<>%
#'
#' @importFrom stringr str_detect str_replace str_squish
#'
#' @import dplyr ggplot2
#'
#' @export
msavisr <- function(mymsa = NULL, myref = NULL, mypath = NULL, refontop = TRUE, myroi = NULL,
hnon = NULL, hmat = NULL, hroi = NULL, wnon = NULL, wmat = NULL, wroi = NULL,
anon = NULL, amat = NULL, aroi = NULL, basecolors = NULL, roicolors = NULL,
cbfcols = TRUE){
#To avoid R-build error notes "no visible binding for global variable" and "Undefined global functions or variables:"
outcol <- NULL
roicol <- NULL
#Basic checks
if(is.null(mymsa)) { stop("Please supply the name of a MSA file! (Must be fasta formatted!)") }
if(is.null(myref)) { stop("Please indicate which sequence you would like to use as the reference!") }
#Setting up the path to the object
if(!is.null(mypath)){
mymsa <- paste0(mypath, "/", mymsa)
}
#Reading in the fasta data as a newline-delimited data.frame
df <- utils::read.table(mymsa, sep = "\n")
#Container data.frame where all the computations will occur
fasdf <- data.frame(curhead = c(), curseq = c(), stringsAsFactors = FALSE)
#Identifying those rows that contain the fasta headers
headerpos <- which(stringr::str_detect(df$V1, "^>"))
#Using headerpos and a for loop to create a two column data.frame
#wherein each column in each row represents the fasta header and fasta sequence respectively
for(i in 1:length(headerpos)){
if(i == length(headerpos)){
curpos <- headerpos[i]
nextpos <- nrow(df)
curhead <- df$V1[curpos]
curseq <- paste0(df$V1[(curpos+1):nextpos], collapse = "")
} else{
curpos <- headerpos[i]
nextpos <- headerpos[i+1]
curhead <- df$V1[curpos]
curseq <- paste0(df$V1[(curpos+1):(nextpos-1)], collapse = "")
}
fasdf <- dplyr::bind_rows(fasdf, data.frame(curhead, curseq, stringsAsFactors = FALSE))
if(i == length(headerpos)){
rm(curpos, nextpos, curhead, curseq, headerpos)
}
}
rm(i)
#Need a column that will contain the length of each sequence
#Since this is a MSA--all sequences are of the same length, so length of just one of the is sufficient
curlen <- nchar(fasdf$curseq[1])
#Moving each character of each sequence into its own row
fasdf %<>% dplyr::mutate(curseq = stringr::str_squish(curseq))
fasdf %<>% tidyr::separate_rows(curseq, sep = "")
#separate_rows() adds an additional column for whatever reason (probably because "" is used as the separator)
#Removing this; however, the sequence length calculated above is the "correct" one, so removing this, does not affect
#that (this blank is in addition)
fasdf %<>% dplyr::filter(curseq != "")
#Using the curlen variable above, creating a position column representing each character's position
#in the sequence
fasdf %<>% dplyr::mutate(curpos = rep(1:curlen, len = nrow(fasdf)))
#Cleaning the > in the fasta header column
fasdf %<>% dplyr::mutate(curhead = stringr::str_replace(curhead, "^>", ""))
#The user can select whether the reference sequence goes on the top or bottom of the plot
olvls <- fasdf %>% dplyr::filter(curhead != myref) %>% dplyr::distinct(curhead)
#olvls <- as.character(olvls)
#refontop <- FALSE
if(isTRUE(refontop)){
olvls <- rbind(olvls, myref)
} else{
olvls <- rbind(myref, olvls)
}
olvls <- olvls$curhead
#Reordering
fasdf$curhead <- factor(fasdf$curhead, levels = olvls)
rm(olvls)
#To visualize SNPs and gaps, some one sequence needs to be set as the reference
#This is passed to the function
#First taking this reference sequence out separately and replicate()'ing it vector
#of length (num_sequences) * (num_chars_in_seq); the idea is to cbind() this to the entire data.frame
#and then mark each character as a perfect match (*), gap (-), or SNP (+) by simply comparing
#the sequence column (curseq) against this reference column (refcol)
#That's what's being done below
myrefseq <- fasdf %>% dplyr::filter (curhead == myref) %>% dplyr::select(curseq)
refcol <- unlist(replicate(length(unique(fasdf$curhead)), myrefseq$curseq, simplify = FALSE))
fasdf <- cbind(fasdf, refcol)
rm(myrefseq, refcol)
#fasdf$outcol <- NA
for(i in 1:nrow(fasdf)){
if(fasdf$curseq[i] == fasdf$refcol[i]){
fasdf$outcol[i] <- "Match" #"*" Identical
} else if(fasdf$curseq[i] != fasdf$refcol[i] & fasdf$curseq[i] == "-"){
fasdf$outcol[i] <- "Gap" #"-" Gap
} else{
fasdf$outcol[i] <- "Mismatch" #"+" SNP
}
}
#The user can also manually specify regions of interest (ROI; i.e., positions in the MSA)
#that they wish to highlight manually via myroi as a list of vectors wherein each vector
#is of the format c(seqname, pos, desc), or c(seqname, pos), or c(pos, desc), or c(pos).
#The seqname (name of a specific sequence in the MSA) and the desc (description of the feature)
#are optional. pos (i.e., the positions that denote the ROI) can be a single integer (e.g., 100),
#or a range of integers (1:10). Each such ROI vector can also hold one or more (and combinations) of
#single integer pos values and integer range pos values (e.g., something like c("Seq1", 1:10, 100, "Helix")).
#If such are specified, they also need to be indicated in fasdf, and additonal legend
#items and such must be set up, which is what is being done in the if block below along with the
#concomitant plotting.
#Checking if regions of interest have been supplied and plot accordingly
#If they have been supplied
if(!is.null(myroi)){
#Create an empty column that will indicate whether the particular position
#is a region of interest or not
fasdf$roicol <- NA
#Iterate through the ROIs and indicate them as ROIs in fasdf's outcol
#Get the items in the current position in the list
for(i in 1:length(myroi)){
curset <- myroi[[i]]
curset
#Interpreting the structure of the ROIs depends upon whether or not the first and
#last items in the vector are alphanumeric
lind <- length(curset)
hasroiname <- ifelse(stringr::str_detect(curset[1], "[A-Za-z]+"), TRUE, FALSE)
hasroidesc <- ifelse(stringr::str_detect(curset[lind], "[A-Za-z]+"), TRUE, FALSE)
#The user can supply specific positions in a specific sequence, or specific positions
#universal to all sequences
#To check for this, need to first see if the very first item in curset is a character
#string or not, and proceed from there
#If a specific sequence, positions, and description have been provided
if(hasroiname & hasroidesc){
roiseq <- curset[1]
roipos <- curset[-1]
roipos <- roipos[-length(roipos)]
roidesc <- curset[length(curset)]
roiidx <- paste0(roipos[1], "-", roipos[length(roipos)])
fasdf %<>% dplyr::mutate(roicol = ifelse(curhead == roiseq & curpos %in% roipos,
paste0("ROI ", roiseq, " ", roidesc, " ", roiidx), roicol))
} else if(hasroiname & !hasroidesc){
#If a specific sequence and positions have been provided but no description
roiseq <- curset[1]
roipos <- curset[-1]
roiidx <- paste0(roipos[1], "-", roipos[length(roipos)])
fasdf %<>% dplyr::mutate(roicol = ifelse(curhead == roiseq & curpos %in% roipos,
paste0("ROI ", roiseq, " ", outcol, " ", roiidx), roicol))
} else if(!hasroiname & hasroidesc){
#If positions and a descriptions have been provided, but no specific sequence
roipos <- curset[-length(curset)]
roidesc <- curset[length(curset)]
roiidx <- paste0(roipos[1], "-", roipos[length(roipos)])
fasdf %<>% dplyr::mutate(roicol = ifelse(curpos %in% roipos,
paste0("ROI ", roidesc, " ", roiidx), roicol))
} else if(!hasroiname & !hasroidesc){
#If only positions have been provided
roiidx <- paste0(roipos[1], "-", roipos[length(roipos)])
fasdf %<>% dplyr::mutate(roicol = ifelse(curpos %in% roipos,
paste0("ROI ", outcol, " ", roiidx), roicol))
}
}
#rm(roidesc, roipos, roiseq, lind, i, hasroiname, hasroidesc, curset)
#For all remaining roicol positions that are just NA, fill them with their equivalent
#outcol values
#
fasdf %<>% dplyr::mutate(roicol = ifelse(is.na(roicol), outcol, roicol))
#unique(fasdf$roicol)
#For ordering the legend items - roilvls are ordered in ascending alphabetical order
baselvls <- c("Match", "Mismatch", "Gap")
roilvls <- unique(fasdf$roicol)[!(unique(fasdf$roicol) %in% baselvls)]
roilvls <- roilvls[order(roilvls)]
fasdf$roicol <- factor(fasdf$roicol, levels = c(baselvls, roilvls))
#fasdf$outcol <- factor(fasdf$outcol, levels = c("Match", "Mismatch", "Gap"))
#Setting up the colors for plotting
#The color palette consists of colors for the basic match, mismatch, gap trio
#plus whatever are the n categories of ROIs
#Setting up the colors for plotting
#The color palette consists of colors for the basic match, mismatch, gap trio
#plus whatever are the n categories of ROIs
cbPalette <- c()
#There are two color sets to choose: basecolors and roicolors
#Basecolors
if(!is.null(basecolors)){ #User-supplied basecolors
#Check if too many colors have been supplied
if(length(basecolors) > length(baselvls)){
warning("ROI-mode: Too many basecolors supplied, will be leaving out the last few!!")
basecolors <- basecolors[1:length(baselvls)]
}
#Check if enough colors has been supplied, else choose additional ones automatically
if(length(basecolors) < length(baselvls)){
warning("ROI-mode: Not enough basecolors supplied, choosing some random ones!!")
mydiff <- abs(length(baselvls) - length(basecolors))
repeat{
if(isTRUE(cbfcols)){
extracolors <- sample(unique(viridis::viridis(100)), mydiff)
} else{
extracolors <- sample(grDevices::colors(distinct = TRUE), mydiff)
}
if(!any(extracolors %in% basecolors)) { break }
}
basecolors <- c(basecolors, extracolors)
}
} else{#No user-supplied basecolors
#Check if cbfcols is set and assign appropriately
if(isTRUE(cbfcols)){
basecolors <- sample(unique(viridis::viridis(100)), length(baselvls))
} else{
basecolors <- c("gray", "black", "white")
#basecolors <- sample(grDevices::colors(distinct = TRUE), length(baselvls))
}
}
#roicolors
if(!is.null(roicolors)){ #User-supplied roicolors
#Check if too many colors have been supplied
if(length(roicolors) > length(roilvls)){
warning("ROI-mode: Too many roicolors supplied, will be leaving out the last few!!")
roicolors <- roicolors[1:length(roilvls)]
}
#Check if enough colors has been supplied, else add colors automatically
if(length(roicolors) < length(roilvls)){
warning("ROI-mode: Not enough roicolors supplied, choosing some random ones!!")
mydiff <- abs(length(roilvls) - length(roicolors))
repeat{
if(isTRUE(cbfcols)){
extracolors <- sample(unique(viridis::viridis(100)), mydiff)
} else{
extracolors <- sample(grDevices::colors(distinct = TRUE), mydiff)
}
#Make sure the colors aren't ones that are already picked!!
if(!any(extracolors %in% c(roicolors, basecolors))) { break }
}
roicolors <- c(roicolors, extracolors)
}
} else{
#Check if cbfcols is set and assign appropriately
if(isTRUE(cbfcols)){
repeat{ #To ensure that the basecolors and roicolors aren't the same
roicolors <- sample(unique(viridis::viridis(100)), length(roilvls))
if(!any(roicolors %in% basecolors)) { break }
}
} else{
repeat{ #To ensure that the basecolors and roicolors aren't the same
roicolors <- sample(grDevices::colors(distinct = TRUE), length(roilvls))
if(!any(roicolors %in% basecolors)) { break }
}
}
}
cbPalette <- c(basecolors, roicolors)
#rm(baselvls, roilvls)
#For plotting, the matches, nonmatches (gaps + mismatches), and ROIs need to be
#supplied as separate layers
matches <- fasdf %>% dplyr::filter(roicol == "Match")
nonmatches <- fasdf %>% dplyr::filter(roicol != "Match" & !stringr::str_detect(roicol, "^ROI"))
rois <- fasdf %>% dplyr::filter(stringr::str_detect(roicol, "^ROI"))
#To make specific features (e.g. a ROI or gaps) stand out, each layer has its own width and height
#parameters
#Defaults supplied below can be overridden by user inputs
# hmat <- 0.4
# hnon <- 0.8
# hroi <- 0.8
# wmat <- 1.0
# wnon <- 4.0
# wroi <- 4.0
# amat <- 1.0
# anon <- 1.0
# aroi <- 1.0
#Defaults
if(is.null(hmat)) {hmat <- 0.4}
if(is.null(hnon)) {hnon <- 0.4}
if(is.null(hroi)) {hroi <- 0.4}
if(is.null(wmat)) {wmat <- 1.0}
if(is.null(wnon)) {wnon <- 2.0}
if(is.null(wroi)) {wroi <- 4.0}
if(is.null(amat)) {amat <- 1.0}
if(is.null(anon)) {anon <- 1.0}
if(is.null(aroi)) {aroi <- 1.0}
#Plotting
myplt <- ggplot2::ggplot() +
ggplot2::geom_tile(data = matches,
ggplot2::aes(x = curpos, y = curhead, fill = roicol), width = wmat, height = hmat, alpha = amat) +
ggplot2::geom_tile(data = nonmatches,
ggplot2:: aes(x = curpos, y = curhead, fill = roicol), width = wnon, height = hnon, alpha = anon) +
ggplot2::geom_tile(data = rois,
ggplot2::aes(x = curpos, y = curhead, fill = roicol), width = wroi, height = hroi, alpha = aroi) +
ggplot2::xlab("Position in sequence") + ggplot2::ylab("") +
ggplot2::scale_fill_manual(name = "", values = cbPalette, drop = FALSE) +
ggplot2::theme_classic() +
ggplot2::theme(legend.key = ggplot2::element_rect(colour = "black", size = 1.0),
axis.line.y = ggplot2::element_blank(), axis.line.x = ggplot2::element_blank(),
axis.text.y = ggplot2::element_text(face = "bold"), axis.ticks.y = ggplot2::element_blank())
myplt
}
#If not regions of interest have been supplied
if(is.null(myroi)){
#For ordering the legend items
baselvls <- c("Match", "Mismatch", "Gap")
fasdf$outcol <- factor(fasdf$outcol, levels = baselvls)
#Setting up the colors - basecolors only here
cbPalette <- c()
#Basecolors
if(!is.null(basecolors)){ #User-supplied basecolors
#Check if enough colors has been supplied, else add colors automatically
if(length(basecolors) < length(baselvls)){
warning("No-ROI-mode: Not enough basecolors supplied, choosing some random ones!!")
mydiff <- abs(length(baselvls) - length(basecolors))
repeat{
if(isTRUE(cbfcols)){
extracolors <- sample(unique(viridis::viridis(100)), mydiff)
} else{
extracolors <- sample(grDevices::colors(distinct = TRUE), mydiff)
}
if(!any(extracolors %in% basecolors)) { break }
}
}
if(length(basecolors) > length(baselvls)){
warning("No-ROI-mode: Too many basecolors supplied, will be leaving out the last few!!")
basecolors <- basecolors[1:length(baselvls)]
}
} else{#No user-supplied basecolors
#Check if cbfcols is set and assign appropriately
if(isTRUE(cbfcols)){
basecolors <- sample(unique(viridis::viridis(100)), length(baselvls))
} else{
basecolors <- c("gray", "black", "white")
}
}
cbPalette <- basecolors
#Not ROIs so only two layers: matches and non-matches as layers for ggplot()
matches <- fasdf %>% dplyr::filter(outcol == "Match")
nonmatches <- fasdf %>% dplyr::filter(outcol != "Match")
#Width, height, alpha parameters for each geom_tile layer
#Defaults supplied below can be overridden by user inputs
# hmat <- 0.4
# hnon <- 0.8
#hroi <- 0.8
# wmat <- 1.0
# wnon <- 4.0
#wroi <- 4.0
# amat <- 1.0
# anon <- 1.0
#aroi <- 1.0
#Defaults
if(is.null(hmat)) {hmat <- 0.4}
if(is.null(hnon)) {hnon <- 0.4}
if(is.null(hroi)) {hroi <- 0.4}
if(is.null(wmat)) {wmat <- 1.0}
if(is.null(wnon)) {wnon <- 2.0}
if(is.null(wroi)) {wroi <- 4.0}
if(is.null(amat)) {amat <- 1.0}
if(is.null(anon)) {anon <- 1.0}
if(is.null(aroi)) {aroi <- 1.0}
#Plotting
myplt <- ggplot2::ggplot() +
ggplot2::geom_tile(data = matches,
ggplot2::aes(x = curpos, y = curhead, fill = outcol), width = wmat, height = hmat, alpha = amat) +
ggplot2::geom_tile(data = nonmatches,
ggplot2::aes(x = curpos, y = curhead, fill = outcol), width = wnon, height = hnon, alpha = anon) +
ggplot2::xlab("Position in sequence") + ggplot2::ylab("") +
ggplot2::scale_fill_manual(name = "", values = cbPalette, drop = FALSE) +
ggplot2::theme_classic() +
ggplot2::theme(legend.key = ggplot2::element_rect(colour = "black", size = 1.0),
axis.line.y = ggplot2::element_blank(), axis.line.x = ggplot2::element_blank(),
axis.text.y = ggplot2::element_text(face = "bold"), axis.ticks.y = ggplot2::element_blank())
myplt
}
return(myplt)
}
vragh/seqvisr documentation built on April 20, 2024, 10:06 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions msavisr
Documented in msavisr
#'Multiple sequence alignment (MSA) visualization and manipulation thereof.
#'
#'Given a fasta formatted MSA, msavisr will attempt to produce a visualization of the MSA
#'with matches, mismatches, gaps, and (optional) regions of interest (ROIs) being highlighted
#'in different colors. The matches, mismatches, and gaps are calculated by comparing all other
#'sequences to a designated reference sequence within the MSA. ROIs are specified manually (see
#'below).
#'
#'
#' @usage msavisr(mymsa = NULL, myref = NULL, mypath = NULL,
#' refontop = TRUE, myroi = NULL, hnon = NULL, hmat = NULL,
#' hroi = NULL, wnon = NULL, wmat = NULL, wroi = NULL,
#' anon = NULL, amat = NULL, aroi = NULL, basecolors = NULL,
#' roicolors = NULL, cbfcols = TRUE)
#'
#' @param mymsa (character string, mandatory) the name of the fasta-formatted file containing the
#' multiple sequence alignment (MSA). The full path to the file can also be provided here (in which
#' case, the mypath argument must be set to NULL).
#'
#' @param myref (character string, mandatory) the fasta header (not including the ">") of one of the
#' sequences in the MSA which act as the reference sequence.
#'
#' @param mypath (character string, optional) the path to the directory containing the MSA. (Default
#' : looks in the current directory.)
#'
#' @param refontop (boolean, optional) should the reference sequence be placed at the top of the MSA plot
#' (default) or at the bottom? (Set to FALSE for bottom.)
#'
#' @param myroi (list of vectors, optional) the user can manually specify regions of interest (ROI;
#' i.e., positions in the MSA) that they wish to highlight manually via myroi. This must be supplied
#' as a list of vectors wherein each vector is of the format c(seqname, pos, desc), or c(seqname, pos),
#' or c(pos, desc), or c(pos). The seqname (name of a specific sequence in the MSA) and the desc
#' (description of the feature) are optional. pos (i.e., the positions that denote the ROI) can be a
#' single integer (e.g., 100), or a range of integers (1:10). Each such ROI vector can also hold one
#' or more (and combinations) of single integer pos values and integer range pos values (e.g., something
#' like c("Seq1", 1:10, 100, "Helix")). Thus, for example, if a SNP at position 100 in "seq2" and a the
#' region of nucleotides/amino-acids 20:30 representing a domain in all sequences were to be ROIs, they
#' would be supplied like so list(c("seq2", 100, "SNP"), c(20:30, "domain")). ROIs can be assigned any
#' color by the user (see argument roicolors); colors are assigned automatically otherwise (optionally
#' from a color blind-friendly palette; see argument cbfcols).
#'
#' @param hnon (double, optional) the height of the feature "bar(s)" for all mismatches and gaps in the MSA. (Default: 0.4.)
#'
#' @param hmat (double, optional) the height of the feature "bar(s)" for all matches in the MSA. (Default: 0.4.)
#'
#' @param hroi (double, optional) the height of the feature "bar(s)" for all ROIs in the MSA. (Default: 0.4.)
#'
#' @param wnon (double, optional) the width of the feature "bar(s)" for all mismatches and gaps in the MSA. (Default: 2.0.)
#'
#' @param wmat (double, optional) the width of the feature "bar(s)" for all matches in the MSA. (Default: 1.0.)
#'
#' @param wroi (double, optional) the width of the feature "bar(s)" for all ROIs in the MSA. (Default: 4.0.)
#'
#' @param anon (double, optional) the transparency of the feature "bar(s)" for all mismatches and gaps in the MSA. (Default: 1.0.)
#'
#' @param amat (double, optional) the transparency of the feature "bar(s)" for all matches in the MSA. (Default: 1.0.)
#'
#' @param aroi (double, optional) the transparency of the feature "bar(s)" for all ROIs in the MSA. (Default: 1.0.)
#'
#' @param basecolors (vector of 3 character strings, optional) the colors for the matches, mismatches, and gaps can optionally be
#' supplied by the user via this argument. Defaults are c("gray", "black", "white") when no colors are supplied and cbfcols (see
#' below) is set to FALSE. If cbfcols == TRUE, and no colors are supplied by the user then a palette from viridis is chosen.
#'
#' @param roicolors (vector of n character strings, optional) user-specified colors for the ROI features can be supplied via this
#' argument. As many colors as there are ROIs need to be supplied, and the order of the colors should correspond to the order of the
#' ROIs in the input list. If too few colors are supplied, colors are reused; if too many are supplied, the last few colors will not
#' be used. Defaults are chosen automacially from grDevices::color() if cbfcols == FALSE, and from viridis otherwise.
#'
#' @param cbfcols (boolean, optional) allows for the user to choose whether the automatic coloring scheme used should be color-blind
#' friendly. (Default: TRUE; set to FALSE to use non-color blind-friendly colors.)
#'
#' @details
#' msavisr plots the matches as a separate geom_tile() layer, the gaps + mismatches as a geom_tile() layer, and the ROIs as a
#' separate geom_tile() layer (if ROIs are supplied).
#'
#' The user will have to tweak the values for the widths and heights of these layers (via the arguments outlined above) to achieve
#' the desired visualization "effects". In general, it is advisable to set the widths of the mismatches + gaps (and/or ROIs, if any;
#' so wnon and wroi respectively) larger than that of the matches (wmat). The heights can be increased if necessary. Altering the
#' transparency levels does not seem to be very useful. Note: altering the transparency levels does not update the transparency of
#' the colors shown in the legend!!
#'
#' ROIs are especially useful for visualizing features such as single nucleotide polymorphisms (SNPs) in nucleotide MSAs and other
#' such isolated features that might normally become "buried" within the bulk of the sequence. This can be easily achieved by
#' indicating the SNPs position as an ROI and jacking up its width (wroi) and/or height (hroi) values.
#'
#' @return A ggplot2 object is returned to the parent environment for plotting and/or further downstream processing/manipulation.
#'
#' @note
#' The only issue with specifying ROIs in the manner implemented here is that, if for instance, a ROI in "seq2" at position 100
#' and in "seq4" at the same position need to be highlighted, it cannot be supplied like so c("seq2", "seq4", 100), and instead must
#' be supplied as two separate vectors c("seq2", 100), c("seq4", 100). The lowdown: specifying common ROIs in a SUBSET of the MSA can
#' be a tedious process. Unfortunately, as of now, no internal workarounds have been implemented.
#'
#' @examples
#' \dontrun{
#' #Input data
#' testmsa <- system.file("extdata", "testaln_mrna.fasta", package = "seqvisr", mustWork = TRUE)
#'
#' #Basic visualization
#' msavisr(mymsa = testmsa, myref = "Ref0") #No ROIs
#'
#' #Defining an example ROI
#' testroi <- list(c("Ref0", 100:110, "Ref0 Domain1"), c(14, "Pseudouridine"),
#' c(20:30, "Domain2"), c("Seq2", 55, "SNP"))
#'
#' #MSA with ROIs
#' msavisr(mymsa = testmsa, myref = "Ref0", myroi = testroi)
#' }
#'
#' @importFrom grDevices colors
#'
#' @importFrom utils read.table
#'
#' @importFrom viridis viridis
#'
#' @importFrom tidyr separate_rows
#'
#' @importFrom magrittr %>% %<>%
#'
#' @importFrom stringr str_detect str_replace str_squish
#'
#' @import dplyr ggplot2
#'
#' @export
msavisr <- function(mymsa = NULL, myref = NULL, mypath = NULL, refontop = TRUE, myroi = NULL,
hnon = NULL, hmat = NULL, hroi = NULL, wnon = NULL, wmat = NULL, wroi = NULL,
anon = NULL, amat = NULL, aroi = NULL, basecolors = NULL, roicolors = NULL,
cbfcols = TRUE){
#To avoid R-build error notes "no visible binding for global variable" and "Undefined global functions or variables:"
outcol <- NULL
roicol <- NULL
#Basic checks
if(is.null(mymsa)) { stop("Please supply the name of a MSA file! (Must be fasta formatted!)") }
if(is.null(myref)) { stop("Please indicate which sequence you would like to use as the reference!") }
#Setting up the path to the object
if(!is.null(mypath)){
mymsa <- paste0(mypath, "/", mymsa)
}
#Reading in the fasta data as a newline-delimited data.frame
df <- utils::read.table(mymsa, sep = "\n")
#Container data.frame where all the computations will occur
fasdf <- data.frame(curhead = c(), curseq = c(), stringsAsFactors = FALSE)
#Identifying those rows that contain the fasta headers
headerpos <- which(stringr::str_detect(df$V1, "^>"))
#Using headerpos and a for loop to create a two column data.frame
#wherein each column in each row represents the fasta header and fasta sequence respectively
for(i in 1:length(headerpos)){
if(i == length(headerpos)){
curpos <- headerpos[i]
nextpos <- nrow(df)
curhead <- df$V1[curpos]
curseq <- paste0(df$V1[(curpos+1):nextpos], collapse = "")
} else{
curpos <- headerpos[i]
nextpos <- headerpos[i+1]
curhead <- df$V1[curpos]
curseq <- paste0(df$V1[(curpos+1):(nextpos-1)], collapse = "")
}
fasdf <- dplyr::bind_rows(fasdf, data.frame(curhead, curseq, stringsAsFactors = FALSE))
if(i == length(headerpos)){
rm(curpos, nextpos, curhead, curseq, headerpos)
}
}
rm(i)
#Need a column that will contain the length of each sequence
#Since this is a MSA--all sequences are of the same length, so length of just one of the is sufficient
curlen <- nchar(fasdf$curseq[1])
#Moving each character of each sequence into its own row
fasdf %<>% dplyr::mutate(curseq = stringr::str_squish(curseq))
fasdf %<>% tidyr::separate_rows(curseq, sep = "")
#separate_rows() adds an additional column for whatever reason (probably because "" is used as the separator)
#Removing this; however, the sequence length calculated above is the "correct" one, so removing this, does not affect
#that (this blank is in addition)
fasdf %<>% dplyr::filter(curseq != "")
#Using the curlen variable above, creating a position column representing each character's position
#in the sequence
fasdf %<>% dplyr::mutate(curpos = rep(1:curlen, len = nrow(fasdf)))
#Cleaning the > in the fasta header column
fasdf %<>% dplyr::mutate(curhead = stringr::str_replace(curhead, "^>", ""))
#The user can select whether the reference sequence goes on the top or bottom of the plot
olvls <- fasdf %>% dplyr::filter(curhead != myref) %>% dplyr::distinct(curhead)
#olvls <- as.character(olvls)
#refontop <- FALSE
if(isTRUE(refontop)){
olvls <- rbind(olvls, myref)
} else{
olvls <- rbind(myref, olvls)
}
olvls <- olvls$curhead
#Reordering
fasdf$curhead <- factor(fasdf$curhead, levels = olvls)
rm(olvls)
#To visualize SNPs and gaps, some one sequence needs to be set as the reference
#This is passed to the function
#First taking this reference sequence out separately and replicate()'ing it vector
#of length (num_sequences) * (num_chars_in_seq); the idea is to cbind() this to the entire data.frame
#and then mark each character as a perfect match (*), gap (-), or SNP (+) by simply comparing
#the sequence column (curseq) against this reference column (refcol)
#That's what's being done below
myrefseq <- fasdf %>% dplyr::filter (curhead == myref) %>% dplyr::select(curseq)
refcol <- unlist(replicate(length(unique(fasdf$curhead)), myrefseq$curseq, simplify = FALSE))
fasdf <- cbind(fasdf, refcol)
rm(myrefseq, refcol)
#fasdf$outcol <- NA
for(i in 1:nrow(fasdf)){
if(fasdf$curseq[i] == fasdf$refcol[i]){
fasdf$outcol[i] <- "Match" #"*" Identical
} else if(fasdf$curseq[i] != fasdf$refcol[i] & fasdf$curseq[i] == "-"){
fasdf$outcol[i] <- "Gap" #"-" Gap
} else{
fasdf$outcol[i] <- "Mismatch" #"+" SNP
}
}
#The user can also manually specify regions of interest (ROI; i.e., positions in the MSA)
#that they wish to highlight manually via myroi as a list of vectors wherein each vector
#is of the format c(seqname, pos, desc), or c(seqname, pos), or c(pos, desc), or c(pos).
#The seqname (name of a specific sequence in the MSA) and the desc (description of the feature)
#are optional. pos (i.e., the positions that denote the ROI) can be a single integer (e.g., 100),
#or a range of integers (1:10). Each such ROI vector can also hold one or more (and combinations) of
#single integer pos values and integer range pos values (e.g., something like c("Seq1", 1:10, 100, "Helix")).
#If such are specified, they also need to be indicated in fasdf, and additonal legend
#items and such must be set up, which is what is being done in the if block below along with the
#concomitant plotting.
#Checking if regions of interest have been supplied and plot accordingly
#If they have been supplied
if(!is.null(myroi)){
#Create an empty column that will indicate whether the particular position
#is a region of interest or not
fasdf$roicol <- NA
#Iterate through the ROIs and indicate them as ROIs in fasdf's outcol
#Get the items in the current position in the list
for(i in 1:length(myroi)){
curset <- myroi[[i]]
curset
#Interpreting the structure of the ROIs depends upon whether or not the first and
#last items in the vector are alphanumeric
lind <- length(curset)
hasroiname <- ifelse(stringr::str_detect(curset[1], "[A-Za-z]+"), TRUE, FALSE)
hasroidesc <- ifelse(stringr::str_detect(curset[lind], "[A-Za-z]+"), TRUE, FALSE)
#The user can supply specific positions in a specific sequence, or specific positions
#universal to all sequences
#To check for this, need to first see if the very first item in curset is a character
#string or not, and proceed from there
#If a specific sequence, positions, and description have been provided
if(hasroiname & hasroidesc){
roiseq <- curset[1]
roipos <- curset[-1]
roipos <- roipos[-length(roipos)]
roidesc <- curset[length(curset)]
roiidx <- paste0(roipos[1], "-", roipos[length(roipos)])
fasdf %<>% dplyr::mutate(roicol = ifelse(curhead == roiseq & curpos %in% roipos,
paste0("ROI ", roiseq, " ", roidesc, " ", roiidx), roicol))
} else if(hasroiname & !hasroidesc){
#If a specific sequence and positions have been provided but no description
roiseq <- curset[1]
roipos <- curset[-1]
roiidx <- paste0(roipos[1], "-", roipos[length(roipos)])
fasdf %<>% dplyr::mutate(roicol = ifelse(curhead == roiseq & curpos %in% roipos,
paste0("ROI ", roiseq, " ", outcol, " ", roiidx), roicol))
} else if(!hasroiname & hasroidesc){
#If positions and a descriptions have been provided, but no specific sequence
roipos <- curset[-length(curset)]
roidesc <- curset[length(curset)]
roiidx <- paste0(roipos[1], "-", roipos[length(roipos)])
fasdf %<>% dplyr::mutate(roicol = ifelse(curpos %in% roipos,
paste0("ROI ", roidesc, " ", roiidx), roicol))
} else if(!hasroiname & !hasroidesc){
#If only positions have been provided
roiidx <- paste0(roipos[1], "-", roipos[length(roipos)])
fasdf %<>% dplyr::mutate(roicol = ifelse(curpos %in% roipos,
paste0("ROI ", outcol, " ", roiidx), roicol))
}
}
#rm(roidesc, roipos, roiseq, lind, i, hasroiname, hasroidesc, curset)
#For all remaining roicol positions that are just NA, fill them with their equivalent
#outcol values
#
fasdf %<>% dplyr::mutate(roicol = ifelse(is.na(roicol), outcol, roicol))
#unique(fasdf$roicol)
#For ordering the legend items - roilvls are ordered in ascending alphabetical order
baselvls <- c("Match", "Mismatch", "Gap")
roilvls <- unique(fasdf$roicol)[!(unique(fasdf$roicol) %in% baselvls)]
roilvls <- roilvls[order(roilvls)]
fasdf$roicol <- factor(fasdf$roicol, levels = c(baselvls, roilvls))
#fasdf$outcol <- factor(fasdf$outcol, levels = c("Match", "Mismatch", "Gap"))
#Setting up the colors for plotting
#The color palette consists of colors for the basic match, mismatch, gap trio
#plus whatever are the n categories of ROIs
#Setting up the colors for plotting
#The color palette consists of colors for the basic match, mismatch, gap trio
#plus whatever are the n categories of ROIs
cbPalette <- c()
#There are two color sets to choose: basecolors and roicolors
#Basecolors
if(!is.null(basecolors)){ #User-supplied basecolors
#Check if too many colors have been supplied
if(length(basecolors) > length(baselvls)){
warning("ROI-mode: Too many basecolors supplied, will be leaving out the last few!!")
basecolors <- basecolors[1:length(baselvls)]
}
#Check if enough colors has been supplied, else choose additional ones automatically
if(length(basecolors) < length(baselvls)){
warning("ROI-mode: Not enough basecolors supplied, choosing some random ones!!")
mydiff <- abs(length(baselvls) - length(basecolors))
repeat{
if(isTRUE(cbfcols)){
extracolors <- sample(unique(viridis::viridis(100)), mydiff)
} else{
extracolors <- sample(grDevices::colors(distinct = TRUE), mydiff)
}
if(!any(extracolors %in% basecolors)) { break }
}
basecolors <- c(basecolors, extracolors)
}
} else{#No user-supplied basecolors
#Check if cbfcols is set and assign appropriately
if(isTRUE(cbfcols)){
basecolors <- sample(unique(viridis::viridis(100)), length(baselvls))
} else{
basecolors <- c("gray", "black", "white")
#basecolors <- sample(grDevices::colors(distinct = TRUE), length(baselvls))
}
}
#roicolors
if(!is.null(roicolors)){ #User-supplied roicolors
#Check if too many colors have been supplied
if(length(roicolors) > length(roilvls)){
warning("ROI-mode: Too many roicolors supplied, will be leaving out the last few!!")
roicolors <- roicolors[1:length(roilvls)]
}
#Check if enough colors has been supplied, else add colors automatically
if(length(roicolors) < length(roilvls)){
warning("ROI-mode: Not enough roicolors supplied, choosing some random ones!!")
mydiff <- abs(length(roilvls) - length(roicolors))
repeat{
if(isTRUE(cbfcols)){
extracolors <- sample(unique(viridis::viridis(100)), mydiff)
} else{
extracolors <- sample(grDevices::colors(distinct = TRUE), mydiff)
}
#Make sure the colors aren't ones that are already picked!!
if(!any(extracolors %in% c(roicolors, basecolors))) { break }
}
roicolors <- c(roicolors, extracolors)
}
} else{
#Check if cbfcols is set and assign appropriately
if(isTRUE(cbfcols)){
repeat{ #To ensure that the basecolors and roicolors aren't the same
roicolors <- sample(unique(viridis::viridis(100)), length(roilvls))
if(!any(roicolors %in% basecolors)) { break }
}
} else{
repeat{ #To ensure that the basecolors and roicolors aren't the same
roicolors <- sample(grDevices::colors(distinct = TRUE), length(roilvls))
if(!any(roicolors %in% basecolors)) { break }
}
}
}
cbPalette <- c(basecolors, roicolors)
#rm(baselvls, roilvls)
#For plotting, the matches, nonmatches (gaps + mismatches), and ROIs need to be
#supplied as separate layers
matches <- fasdf %>% dplyr::filter(roicol == "Match")
nonmatches <- fasdf %>% dplyr::filter(roicol != "Match" & !stringr::str_detect(roicol, "^ROI"))
rois <- fasdf %>% dplyr::filter(stringr::str_detect(roicol, "^ROI"))
#To make specific features (e.g. a ROI or gaps) stand out, each layer has its own width and height
#parameters
#Defaults supplied below can be overridden by user inputs
# hmat <- 0.4
# hnon <- 0.8
# hroi <- 0.8
# wmat <- 1.0
# wnon <- 4.0
# wroi <- 4.0
# amat <- 1.0
# anon <- 1.0
# aroi <- 1.0
#Defaults
if(is.null(hmat)) {hmat <- 0.4}
if(is.null(hnon)) {hnon <- 0.4}
if(is.null(hroi)) {hroi <- 0.4}
if(is.null(wmat)) {wmat <- 1.0}
if(is.null(wnon)) {wnon <- 2.0}
if(is.null(wroi)) {wroi <- 4.0}
if(is.null(amat)) {amat <- 1.0}
if(is.null(anon)) {anon <- 1.0}
if(is.null(aroi)) {aroi <- 1.0}
#Plotting
myplt <- ggplot2::ggplot() +
ggplot2::geom_tile(data = matches,
ggplot2::aes(x = curpos, y = curhead, fill = roicol), width = wmat, height = hmat, alpha = amat) +
ggplot2::geom_tile(data = nonmatches,
ggplot2:: aes(x = curpos, y = curhead, fill = roicol), width = wnon, height = hnon, alpha = anon) +
ggplot2::geom_tile(data = rois,
ggplot2::aes(x = curpos, y = curhead, fill = roicol), width = wroi, height = hroi, alpha = aroi) +
ggplot2::xlab("Position in sequence") + ggplot2::ylab("") +
ggplot2::scale_fill_manual(name = "", values = cbPalette, drop = FALSE) +
ggplot2::theme_classic() +
ggplot2::theme(legend.key = ggplot2::element_rect(colour = "black", size = 1.0),
axis.line.y = ggplot2::element_blank(), axis.line.x = ggplot2::element_blank(),
axis.text.y = ggplot2::element_text(face = "bold"), axis.ticks.y = ggplot2::element_blank())
myplt
}
#If not regions of interest have been supplied
if(is.null(myroi)){
#For ordering the legend items
baselvls <- c("Match", "Mismatch", "Gap")
fasdf$outcol <- factor(fasdf$outcol, levels = baselvls)
#Setting up the colors - basecolors only here
cbPalette <- c()
#Basecolors
if(!is.null(basecolors)){ #User-supplied basecolors
#Check if enough colors has been supplied, else add colors automatically
if(length(basecolors) < length(baselvls)){
warning("No-ROI-mode: Not enough basecolors supplied, choosing some random ones!!")
mydiff <- abs(length(baselvls) - length(basecolors))
repeat{
if(isTRUE(cbfcols)){
extracolors <- sample(unique(viridis::viridis(100)), mydiff)
} else{
extracolors <- sample(grDevices::colors(distinct = TRUE), mydiff)
}
if(!any(extracolors %in% basecolors)) { break }
}
}
if(length(basecolors) > length(baselvls)){
warning("No-ROI-mode: Too many basecolors supplied, will be leaving out the last few!!")
basecolors <- basecolors[1:length(baselvls)]
}
} else{#No user-supplied basecolors
#Check if cbfcols is set and assign appropriately
if(isTRUE(cbfcols)){
basecolors <- sample(unique(viridis::viridis(100)), length(baselvls))
} else{
basecolors <- c("gray", "black", "white")
}
}
cbPalette <- basecolors
#Not ROIs so only two layers: matches and non-matches as layers for ggplot()
matches <- fasdf %>% dplyr::filter(outcol == "Match")
nonmatches <- fasdf %>% dplyr::filter(outcol != "Match")
#Width, height, alpha parameters for each geom_tile layer
#Defaults supplied below can be overridden by user inputs
# hmat <- 0.4
# hnon <- 0.8
#hroi <- 0.8
# wmat <- 1.0
# wnon <- 4.0
#wroi <- 4.0
# amat <- 1.0
# anon <- 1.0
#aroi <- 1.0
#Defaults
if(is.null(hmat)) {hmat <- 0.4}
if(is.null(hnon)) {hnon <- 0.4}
if(is.null(hroi)) {hroi <- 0.4}
if(is.null(wmat)) {wmat <- 1.0}
if(is.null(wnon)) {wnon <- 2.0}
if(is.null(wroi)) {wroi <- 4.0}
if(is.null(amat)) {amat <- 1.0}
if(is.null(anon)) {anon <- 1.0}
if(is.null(aroi)) {aroi <- 1.0}
#Plotting
myplt <- ggplot2::ggplot() +
ggplot2::geom_tile(data = matches,
ggplot2::aes(x = curpos, y = curhead, fill = outcol), width = wmat, height = hmat, alpha = amat) +
ggplot2::geom_tile(data = nonmatches,
ggplot2::aes(x = curpos, y = curhead, fill = outcol), width = wnon, height = hnon, alpha = anon) +
ggplot2::xlab("Position in sequence") + ggplot2::ylab("") +
ggplot2::scale_fill_manual(name = "", values = cbPalette, drop = FALSE) +
ggplot2::theme_classic() +
ggplot2::theme(legend.key = ggplot2::element_rect(colour = "black", size = 1.0),
axis.line.y = ggplot2::element_blank(), axis.line.x = ggplot2::element_blank(),
axis.text.y = ggplot2::element_text(face = "bold"), axis.ticks.y = ggplot2::element_blank())
myplt
}
return(myplt)
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.