2020-03-07 21:48:21 +01:00
# ==================================================================== #
# TITLE #
2021-02-02 23:57:35 +01:00
# Antimicrobial Resistance (AMR) Data Analysis for R #
2020-03-07 21:48:21 +01:00
# #
# SOURCE #
2020-07-08 14:48:06 +02:00
# https://github.com/msberends/AMR #
2020-03-07 21:48:21 +01:00
# #
# LICENCE #
2020-12-27 00:30:28 +01:00
# (c) 2018-2021 Berends MS, Luz CF et al. #
2020-10-08 11:16:03 +02:00
# Developed at the University of Groningen, the Netherlands, in #
# collaboration with non-profit organisations Certe Medical #
# Diagnostics & Advice, and University Medical Center Groningen. #
2020-03-07 21:48:21 +01:00
# #
# This R package is free software; you can freely use and distribute #
# it for both personal and commercial purposes under the terms of the #
# GNU General Public License version 2.0 (GNU GPL-2), as published by #
# the Free Software Foundation. #
# We created this package for both routine data analysis and academic #
# research and it was publicly released in the hope that it will be #
# useful, but it comes WITHOUT ANY WARRANTY OR LIABILITY. #
2020-10-08 11:16:03 +02:00
# #
# Visit our website for the full manual and a complete tutorial about #
2021-02-02 23:57:35 +01:00
# how to conduct AMR data analysis: https://msberends.github.io/AMR/ #
2020-03-07 21:48:21 +01:00
# ==================================================================== #
#' Principal Component Analysis (for AMR)
#'
2020-03-08 11:18:59 +01:00
#' Performs a principal component analysis (PCA) based on a data set with automatic determination for afterwards plotting the groups and labels, and automatic filtering on only suitable (i.e. non-empty and numeric) variables.
2021-03-07 13:52:39 +01:00
#' @inheritSection lifecycle Stable Lifecycle
2021-05-12 18:15:03 +02:00
#' @param x a [data.frame] containing [numeric] columns
2020-05-16 13:05:47 +02:00
#' @param ... columns of `x` to be selected for PCA, can be unquoted since it supports quasiquotation.
2020-03-07 21:48:21 +01:00
#' @inheritParams stats::prcomp
2020-03-08 11:18:59 +01:00
#' @details The [pca()] function takes a [data.frame] as input and performs the actual PCA with the \R function [prcomp()].
2020-03-07 21:48:21 +01:00
#'
2021-05-12 18:15:03 +02:00
#' The result of the [pca()] function is a [prcomp] object, with an additional attribute `non_numeric_cols` which is a vector with the column names of all columns that do not contain [numeric] values. These are probably the groups and labels, and will be used by [ggplot_pca()].
2020-03-08 11:18:59 +01:00
#' @return An object of classes [pca] and [prcomp]
2020-03-08 09:12:11 +01:00
#' @importFrom stats prcomp
2020-03-07 21:48:21 +01:00
#' @export
2021-01-18 16:57:56 +01:00
#' @inheritSection AMR Read more on Our Website!
2020-03-07 21:48:21 +01:00
#' @examples
2021-01-24 14:48:56 +01:00
#' # `example_isolates` is a data set available in the AMR package.
2020-03-07 21:48:21 +01:00
#' # See ?example_isolates.
#'
2020-09-29 23:35:46 +02:00
#' \donttest{
#'
#' if (require("dplyr")) {
#' # calculate the resistance per group first
#' resistance_data <- example_isolates %>%
#' group_by(order = mo_order(mo), # group on anything, like order
2021-01-28 16:09:30 +01:00
#' genus = mo_genus(mo)) %>% # and genus as we do here;
2020-09-29 23:35:46 +02:00
#' summarise_if(is.rsi, resistance) # then get resistance of all drugs
#'
#' # now conduct PCA for certain antimicrobial agents
#' pca_result <- resistance_data %>%
#' pca(AMC, CXM, CTX, CAZ, GEN, TOB, TMP, SXT)
#'
#' pca_result
#' summary(pca_result)
#' biplot(pca_result)
#' ggplot_pca(pca_result) # a new and convenient plot function
#' }
2020-05-16 13:05:47 +02:00
#' }
2020-03-08 11:18:59 +01:00
pca <- function ( x ,
... ,
retx = TRUE ,
center = TRUE ,
scale. = TRUE ,
tol = NULL ,
rank. = NULL ) {
2020-10-19 17:09:19 +02:00
meet_criteria ( x , allow_class = " data.frame" )
meet_criteria ( retx , allow_class = " logical" , has_length = 1 )
meet_criteria ( center , allow_class = " logical" , has_length = 1 )
meet_criteria ( scale. , allow_class = " logical" , has_length = 1 )
meet_criteria ( tol , allow_class = " numeric" , has_length = 1 , allow_NULL = TRUE )
meet_criteria ( rank. , allow_class = " numeric" , has_length = 1 , allow_NULL = TRUE )
2020-03-07 21:48:21 +01:00
2020-03-08 11:18:59 +01:00
# unset data.table, tibble, etc.
2020-03-07 21:48:21 +01:00
# also removes groups made by dplyr::group_by
x <- as.data.frame ( x , stringsAsFactors = FALSE )
x.bak <- x
2020-05-16 13:05:47 +02:00
# defuse R expressions, this replaces rlang::enquos()
dots <- substitute ( list ( ... ) )
if ( length ( dots ) > 1 ) {
2020-03-07 21:48:21 +01:00
new_list <- list ( 0 )
2020-05-16 13:05:47 +02:00
for ( i in seq_len ( length ( dots ) - 1 ) ) {
new_list [ [i ] ] <- tryCatch ( eval ( dots [ [i + 1 ] ] , envir = x ) ,
2020-03-07 21:48:21 +01:00
error = function ( e ) stop ( e $ message , call. = FALSE ) )
if ( length ( new_list [ [i ] ] ) == 1 ) {
2020-05-16 13:05:47 +02:00
if ( is.character ( new_list [ [i ] ] ) & new_list [ [i ] ] %in% colnames ( x ) ) {
2020-09-18 16:05:53 +02:00
# this is to support quoted variables: df %pm>% pca("mycol1", "mycol2")
2020-05-16 13:05:47 +02:00
new_list [ [i ] ] <- x [ , new_list [ [i ] ] ]
2020-03-07 21:48:21 +01:00
} else {
2020-12-22 00:51:17 +01:00
# remove item - it's a argument like `center`
2020-03-07 21:48:21 +01:00
new_list [ [i ] ] <- NULL
}
}
}
2020-05-16 13:05:47 +02:00
2020-03-07 21:48:21 +01:00
x <- as.data.frame ( new_list , stringsAsFactors = FALSE )
2020-12-28 22:24:33 +01:00
if ( any ( vapply ( FUN.VALUE = logical ( 1 ) , x , function ( y ) ! is.numeric ( y ) ) ) ) {
2021-05-12 18:15:03 +02:00
warning_ ( " Be sure to first calculate the resistance (or susceptibility) of variables with antimicrobial test results, since PCA works with [numeric] variables only. See Examples in ?pca." , call = FALSE )
2020-03-07 21:48:21 +01:00
}
2020-05-16 13:05:47 +02:00
2020-03-07 21:48:21 +01:00
# set column names
2020-05-16 13:05:47 +02:00
tryCatch ( colnames ( x ) <- as.character ( dots ) [2 : length ( dots ) ] ,
2020-03-07 21:48:21 +01:00
error = function ( e ) warning ( " column names could not be set" ) )
2020-05-16 13:05:47 +02:00
2021-05-18 11:29:31 +02:00
# keep only numeric columns
2020-12-28 22:24:33 +01:00
x <- x [ , vapply ( FUN.VALUE = logical ( 1 ) , x , function ( y ) is.numeric ( y ) ) ]
2020-03-07 21:48:21 +01:00
# bind the data set with the non-numeric columns
2020-12-28 22:24:33 +01:00
x <- cbind ( x.bak [ , vapply ( FUN.VALUE = logical ( 1 ) , x.bak , function ( y ) ! is.numeric ( y ) & ! all ( is.na ( y ) ) ) , drop = FALSE ] , x )
2020-03-07 21:48:21 +01:00
}
2020-12-13 20:44:32 +01:00
x <- pm_ungroup ( x ) # would otherwise select the grouping vars
2020-05-16 13:05:47 +02:00
x <- x [rowSums ( is.na ( x ) ) == 0 , ] # remove columns containing NAs
2020-03-08 11:18:59 +01:00
2020-12-28 22:24:33 +01:00
pca_data <- x [ , which ( vapply ( FUN.VALUE = logical ( 1 ) , x , function ( x ) is.numeric ( x ) ) ) ]
2020-03-08 11:18:59 +01:00
2021-02-04 16:48:16 +01:00
message_ ( " Columns selected for PCA: " , vector_and ( font_bold ( colnames ( pca_data ) , collapse = NULL ) , quotes = TRUE ) ,
2020-10-27 15:56:51 +01:00
" . Total observations available: " , nrow ( pca_data ) , " ." )
2020-03-08 11:18:59 +01:00
2021-05-18 11:29:31 +02:00
if ( current_R_older_than ( 3.4 ) ) {
2021-01-28 16:09:30 +01:00
# stats::prcomp prior to 3.4.0 does not have the 'rank.' argument
pca_model <- prcomp ( pca_data , retx = retx , center = center , scale. = scale. , tol = tol )
} else {
pca_model <- prcomp ( pca_data , retx = retx , center = center , scale. = scale. , tol = tol , rank. = rank. )
}
groups <- x [ , vapply ( FUN.VALUE = logical ( 1 ) , x , function ( y ) ! is.numeric ( y ) & ! all ( is.na ( y ) ) ) , drop = FALSE ]
rownames ( groups ) <- NULL
attr ( pca_model , " non_numeric_cols" ) <- groups
2020-03-08 11:18:59 +01:00
class ( pca_model ) <- c ( " pca" , class ( pca_model ) )
pca_model
2020-03-07 21:48:21 +01:00
}
2021-01-28 16:09:30 +01:00
#' @method print pca
#' @export
#' @noRd
print.pca <- function ( x , ... ) {
a <- attributes ( x ) $ non_numeric_cols
if ( ! is.null ( a ) ) {
print_pca_group ( a )
class ( x ) <- class ( x ) [class ( x ) != " pca" ]
}
print ( x , ... )
}
#' @method summary pca
#' @export
#' @noRd
summary.pca <- function ( object , ... ) {
a <- attributes ( object ) $ non_numeric_cols
if ( ! is.null ( a ) ) {
print_pca_group ( a )
class ( object ) <- class ( object ) [class ( object ) != " pca" ]
}
summary ( object , ... )
}
print_pca_group <- function ( a ) {
grps <- sort ( unique ( a [ , 1 , drop = TRUE ] ) )
cat ( " Groups (n=" , length ( grps ) , " , named as '" , colnames ( a ) [1 ] , " '):\n" , sep = " " )
print ( grps )
cat ( " \n" )
}