|
a |
|
b/R/preprocessCELFiles.R |
|
|
1 |
#' RMA preprocess CEL files |
|
|
2 |
#' |
|
|
3 |
#' General function for RMA processing microarray data. Automatically downloads |
|
|
4 |
#' custom Brainarray chip definition files (cdf) if wanted. |
|
|
5 |
#' |
|
|
6 |
#' @param cel_files A character vector of .CEL files. |
|
|
7 |
#' @param cdf A character specifying the CDF type to use. Should be either |
|
|
8 |
#' \code{"brainarray"} or \code{"affy"}. |
|
|
9 |
#' @param target Specifies the "target" CDF to use depending on \code{cdf}. |
|
|
10 |
#' If \code{cdf} is \code{"affy"} then \code{target} is e.g. \code{"core"}. |
|
|
11 |
#' If \code{cdf} is \code{"brainarray"} then \code{target} is |
|
|
12 |
#' e.g. \code{"ensg"}, \code{"entrez"}. |
|
|
13 |
#' See \code{listTargets()} to list the available targets. |
|
|
14 |
#' @param version A character giving the brainarray version to be used. |
|
|
15 |
#' Only used if \code{cdf} is \code{"brainarray"}. See \code{listVersions()}. |
|
|
16 |
#' @param background boolean. Should RMA background correction be performed? |
|
|
17 |
#' @param normalize boolean. Should RMA normalization be performed? |
|
|
18 |
#' @return An expression set object. |
|
|
19 |
#' @importFrom affy read.affybatch just.rma |
|
|
20 |
#' @importFrom affyio read.celfile.header |
|
|
21 |
#' @importFrom oligo rma |
|
|
22 |
#' @export |
|
|
23 |
preprocessCELFiles <- function(cel_files, |
|
|
24 |
cdf = c("affy", "brainarray"), |
|
|
25 |
target, |
|
|
26 |
version = getLatestVersion(), |
|
|
27 |
background = TRUE, |
|
|
28 |
normalize = TRUE) { |
|
|
29 |
getCDF <- function(file) { |
|
|
30 |
return(read.celfile.header(file)$cdfName) |
|
|
31 |
} |
|
|
32 |
array_type <- sapply(cel_files, getCDF) |
|
|
33 |
array_types <- unique(array_type) |
|
|
34 |
if (length(array_types) != 1) { |
|
|
35 |
stop("The CEL files are from more that one array platform. The unique ", |
|
|
36 |
"platforms are: ", paste(array_types, collapse = ", "), |
|
|
37 |
". The arrays should be preprocessed in batches of array platforms.") |
|
|
38 |
} |
|
|
39 |
array_type <- array_type[1] |
|
|
40 |
cel_files <- normalizePath(cel_files) |
|
|
41 |
cdf <- match.arg(cdf) |
|
|
42 |
|
|
|
43 |
if (tolower(cdf) == "affy") { |
|
|
44 |
# Load expression set |
|
|
45 |
es <- oligo::read.celfiles(cel_files) |
|
|
46 |
|
|
|
47 |
# RMA normalize |
|
|
48 |
if (class(es) %in% c("ExonFeatureSet","HTAFeatureSet","GeneFeatureSet")) { |
|
|
49 |
if (missing(target)) stop("No target provided.") |
|
|
50 |
es_rma <- oligo::rma(es, background = background, |
|
|
51 |
normalize = normalize, target = target) |
|
|
52 |
attr(es_rma, "target") <- target |
|
|
53 |
} else { |
|
|
54 |
es_rma <- oligo::rma(es, background = background, |
|
|
55 |
normalize = normalize) |
|
|
56 |
attr(es_rma, "target") <- NULL |
|
|
57 |
} |
|
|
58 |
|
|
|
59 |
} else if (tolower(cdf) == "brainarray") { |
|
|
60 |
|
|
|
61 |
if (missing(target)) stop("No target provided.") |
|
|
62 |
|
|
|
63 |
req <- requireBrainarray(array_type = array_type, |
|
|
64 |
custom_cdf = target, |
|
|
65 |
version = version) |
|
|
66 |
suppressWarnings({ |
|
|
67 |
es_rma <- just.rma(filenames = cel_files, |
|
|
68 |
verbose = TRUE, |
|
|
69 |
cdfname = getCustomCDFName(req$brain_dat, array_type)) |
|
|
70 |
}) |
|
|
71 |
attr(es_rma, "target") <- target |
|
|
72 |
|
|
|
73 |
} else { |
|
|
74 |
stop("cdf == '", cdf, |
|
|
75 |
"' not supported. Should be either 'affy' or 'brainarray'") |
|
|
76 |
} |
|
|
77 |
|
|
|
78 |
# Add attributes |
|
|
79 |
attr(es_rma, "cdf") <- cdf |
|
|
80 |
attr(es_rma, "version") <- version |
|
|
81 |
attr(es_rma, "background") <- background |
|
|
82 |
attr(es_rma, "normalize") <- normalize |
|
|
83 |
|
|
|
84 |
return(es_rma) |
|
|
85 |
} |