annotate plot_distribution.r @ 1:81d08949af15 draft

Uploaded
author ynewton
date Thu, 13 Dec 2012 11:25:25 -0500
parents f91478b63ec6
children cf8d0d54bc78
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
1 #!/usr/bin/Rscript
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
2
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
3 #usage, options and doc goes here
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
4 argspec <- c("normalize.r - takes any flat file and normalizes the rows or the columns using various normalizations (median_shift, mean_shift, t_statistic (z-score), exp_fit, normal_fit, weibull_0.5_fit, weibull_1_fit, weibull_1.5_fit, weibull_5_fit). Requires a single header line and a single cloumn of annotation.
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
5 Usage:
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
6 normalize.r input.tab norm_type norm_by > output.tab
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
7 Example:
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
8 Rscript normalize.r test_matrix.tab median_shift column > output.tab
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
9 Rscript normalize.r test_matrix.tab mean_shift row normals.tab > output.tab
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
10 Options:
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
11 input matrix (annotated by row and column names)
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
12 normalization type; available options:
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
13 median_shift - shifts all values by the median or the row/column if no normals are specified, otherwise shifts by the median of normals
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
14 mean_shift - shifts all values by the mean or the row/column if no normals are specified, otherwise shifts by the mean of normals
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
15 t_statistic - converts all values to z-scores; if normals are specified then converts to z-scores within normal and non-normal classes separately
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
16 exp_fit - (only by column) ranks data and transforms exponential CDF
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
17 normal_fit - (only by column) ranks data and transforms normal CDF
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
18 weibull_0.5_fit - (only by column) ranks data and transforms Weibull CDF with scale parameter = 1 and shape parameter = 0.5
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
19 weibull_1_fit - (only by column) ranks data and transforms Weibull CDF with scale parameter = 1 and shape parameter = 1
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
20 weibull_1.5_fit - (only by column) ranks data and transforms Weibull CDF with scale parameter = 1 and shape parameter = 1.5
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
21 weibull_5_fit - (only by column) ranks data and transforms Weibull CDF with scale parameter = 1 and shape parameter = 5
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
22 normalization by:
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
23 row
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
24 column
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
25 normals_file is an optional parameter which contains a list of column headers from the input matrix, which should be considered as normals
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
26 output file is specified through redirect character >")
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
27
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
28 read_matrix <- function(in_file){
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
29 header <- strsplit(readLines(con=in_file, n=1), "\t")[[1]]
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
30 cl.cols<- 1:length(header) > 1
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
31 data_matrix.df <- read.delim(in_file, header=TRUE, row.names=NULL, stringsAsFactors=FALSE, na.strings="NA", check.names=FALSE)
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
32 data_matrix <- as.matrix(data_matrix.df[,cl.cols])
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
33 rownames(data_matrix) <- data_matrix.df[,1]
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
34 return(data_matrix)
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
35 }
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
36
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
37 main <- function(argv) {
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
38 in_file <- argv[1]
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
39 out_file <- argv[2]
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
40 sink('/dev/null')
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
41
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
42 input_data <- read_matrix(in_file)
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
43
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
44 pdf(out_file, bg="white")
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
45 par(mfrow=c(1,1))
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
46 hist(input_data, col="lightblue", labels=TRUE, main="Histogram", xlab="")
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
47 plot(density(input_data), type="l", col="blue", main="Density")
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
48 dev.off()
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
49 }
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
50
f91478b63ec6 Uploaded
ynewton
parents:
diff changeset
51 main(commandArgs(TRUE))