Mercurial > repos > devteam > dwt_var_perclass

comparison execute_dwt_var_perClass.pl @ 0:27c5c2979e32 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Imported from capsule None
author devteam
date Thu, 23 Jan 2014 12:31:07 -0500
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:27c5c2979e32
1 #!/usr/bin/perl -w
2
3 use warnings;
4 use IO::Handle;
5 use POSIX qw(floor ceil);
6
7 # example: perl execute_dwt_var_perClass.pl hg18_NCNR_10bp_3flanks_deletionHotspot_data_del.txt deletionHotspot 3flanks del
8
9 $usage = "execute_dwt_var_perClass.pl [TABULAR.in] [TABULAR.out] [TABULAR.out] [PDF.out] \n";
10 die $usage unless @ARGV == 4;
11
12 #get the input arguments
13 my $inputFile = $ARGV[0];
14 my $firstOutputFile = $ARGV[1];
15 my $secondOutputFile = $ARGV[2];
16 my $thirdOutputFile = $ARGV[3];
17
18 open (INPUT, "<", $inputFile) || die("Could not open file $inputFile \n");
19 open (OUTPUT1, ">", $firstOutputFile) || die("Could not open file $firstOutputFile \n");
20 open (OUTPUT2, ">", $secondOutputFile) || die("Could not open file $secondOutputFile \n");
21 open (OUTPUT3, ">", $thirdOutputFile) || die("Could not open file $thirdOutputFile \n");
22 open (ERROR, ">", "error.txt") or die ("Could not open file error.txt \n");
23
24 #save all error messages into the error file $errorFile using the error file handle ERROR
25 STDERR -> fdopen( \*ERROR, "w" ) or die ("Could not direct errors to the error file error.txt \n");
26
27 # choosing meaningful names for the output files
28 $max_dwt = $firstOutputFile;
29 $pvalue = $secondOutputFile;
30 $pdf = $thirdOutputFile;
31
32 # count the number of columns in the input file
33 while($buffer = <INPUT>){
34 #if ($buffer =~ m/interval/){
35 chomp($buffer);
36 $buffer =~ s/^#\s*//;
37 @contrl = split(/\t/, $buffer);
38 last;
39 #}
40 }
41 print "The number of columns in the input file is: " . (@contrl) . "\n";
42 print "\n";
43
44 # count the number of motifs in the input file
45 $count = 0;
46 for ($i = 0; $i < @contrl; $i++){
47 $count++;
48 print "# $contrl[$i]\n";
49 }
50 print "The number of motifs in the input file is: $count \n";
51
52 # check if the number of motifs is not a multiple of 12, and round up is so
53 $count2 = ($count/12);
54 if ($count2 =~ m/(\D)/){
55 print "the number of motifs is not a multiple of 12 \n";
56 $count2 = ceil($count2);
57 }
58 else {
59 print "the number of motifs is a multiple of 12 \n";
60 }
61 print "There will be $count2 subfiles\n\n";
62
63 # split infile into subfiles only 12 motif per file for R plotting
64 for ($x = 1; $x <= $count2; $x++){
65 $a = (($x - 1) * 12 + 1);
66 $b = $x * 12;
67
68 if ($x < $count2){
69 print "# data.short $x <- data_test[, +c($a:$b)]; \n";
70 }
71 else{
72 print "# data.short $x <- data_test[, +c($a:ncol(data_test)]; \n";
73 }
74 }
75
76 print "\n";
77 print "There are 4 output files: \n";
78 print "The first output file is a pdf file\n";
79 print "The second output file is a max_dwt file\n";
80 print "The third output file is a pvalues file\n";
81 print "The fourth output file is a test_final_pvalues file\n";
82
83 # write R script
84 $r_script = "get_dwt_varPermut_getMax.r";
85 print "The R file name is: $r_script \n";
86
87 open(Rcmd, ">", "$r_script") or die "Cannot open $r_script \n\n";
88
89 print Rcmd "
90 ######################################################################
91 # plot power spectra, i.e. wavelet variance by class
92 # add code to create null bands by permuting the original data series
93 # get class of maximum significant variance per feature
94 # generate plots and table matrix of variance including p-values
95 ######################################################################
96 library(\"Rwave\");
97 library(\"wavethresh\");
98 library(\"waveslim\");
99
100 options(echo = FALSE)
101
102 # normalize data
103 norm <- function(data){
104 v <- (data-mean(data))/sd(data);
105 if(sum(is.na(v)) >= 1){
106 v<-data;
107 }
108 return(v);
109 }
110
111 dwt_var_permut_getMax <- function(data, names, filter = 4, bc = \"symmetric\", method = \"kendall\", wf = \"haar\", boundary = \"reflection\") {
112 max_var = NULL;
113 matrix = NULL;
114 title = NULL;
115 final_pvalue = NULL;
116 short.levels = NULL;
117 scale = NULL;
118
119 print(names);
120
121 par(mfcol = c(length(names), length(names)), mar = c(0, 0, 0, 0), oma = c(4, 3, 3, 2), xaxt = \"s\", cex = 1, las = 1);
122
123 short.levels <- wd(data[, 1], filter.number = filter, bc = bc)\$nlevels;
124
125 title <- c(\"motif\");
126 for (i in 1:short.levels){
127 title <- c(title, paste(i, \"var\", sep = \"_\"), paste(i, \"pval\", sep = \"_\"), paste(i, \"test\", sep = \"_\"));
128 }
129 print(title);
130
131 # normalize the raw data
132 data<-apply(data,2,norm);
133
134 for(i in 1:length(names)){
135 for(j in 1:length(names)){
136 temp = NULL;
137 results = NULL;
138 wave1.dwt = NULL;
139 out = NULL;
140
141 out <- vector(length = length(title));
142 temp <- vector(length = short.levels);
143
144 if(i < j) {
145 plot(temp, type = \"n\", axes = FALSE, xlab = NA, ylab = NA);
146 box(col = \"grey\");
147 grid(ny = 0, nx = NULL);
148 } else {
149 if (i > j){
150 plot(temp, type = \"n\", axes = FALSE, xlab = NA, ylab = NA);
151 box(col = \"grey\");
152 grid(ny = 0, nx = NULL);
153 } else {
154
155 wave1.dwt <- dwt(data[, i], wf = wf, short.levels, boundary = boundary);
156
157 temp_row = (short.levels + 1 ) * -1;
158 temp_col = 1;
159 temp <- wave.variance(wave1.dwt)[temp_row, temp_col];
160
161 #permutations code :
162 feature1 = NULL;
163 null = NULL;
164 var_25 = NULL;
165 var_975 = NULL;
166 med = NULL;
167
168 feature1 = data[, i];
169 for (k in 1:1000) {
170 nk_1 = NULL;
171 null.levels = NULL;
172 var = NULL;
173 null_wave1 = NULL;
174
175 nk_1 = sample(feature1, length(feature1), replace = FALSE);
176 null.levels <- wd(nk_1, filter.number = filter, bc = bc)\$nlevels;
177 var <- vector(length = length(null.levels));
178 null_wave1 <- dwt(nk_1, wf = wf, short.levels, boundary = boundary);
179 var<- wave.variance(null_wave1)[-8, 1];
180 null= rbind(null, var);
181 }
182 null <- apply(null, 2, sort, na.last = TRUE);
183 var_25 <- null[25, ];
184 var_975 <- null[975, ];
185 med <- (apply(null, 2, median, na.rm = TRUE));
186
187 # plot
188 results <- cbind(temp, var_25, var_975);
189 matplot(results, type = \"b\", pch = \"*\", lty = 1, col = c(1, 2, 2), axes = F);
190
191 # get pvalues by comparison to null distribution
192 out <- (names[i]);
193 for (m in 1:length(temp)){
194 print(paste(\"scale\", m, sep = \" \"));
195 print(paste(\"var\", temp[m], sep = \" \"));
196 print(paste(\"med\", med[m], sep = \" \"));
197 pv = tail = NULL;
198 out <- c(out, format(temp[m], digits = 3));
199 if (temp[m] >= med[m]){
200 # R tail test
201 print(\"R\");
202 tail <- \"R\";
203 pv <- (length(which(null[, m] >= temp[m])))/(length(na.exclude(null[, m])));
204
205 } else {
206 if (temp[m] < med[m]){
207 # L tail test
208 print(\"L\");
209 tail <- \"L\";
210 pv <- (length(which(null[, m] <= temp[m])))/(length(na.exclude(null[, m])));
211 }
212 }
213 out <- c(out, pv);
214 print(pv);
215 out <- c(out, tail);
216 }
217 final_pvalue <-rbind(final_pvalue, out);
218
219
220 # get variances outside null bands by comparing temp to null
221 ## temp stores variance for each scale, and null stores permuted variances for null bands
222 for (n in 1:length(temp)){
223 if (temp[n] <= var_975[n]){
224 temp[n] <- NA;
225 } else {
226 temp[n] <- temp[n];
227 }
228 }
229 matrix <- rbind(matrix, temp)
230 }
231 }
232 # labels
233 if (i == 1){
234 mtext(names[j], side = 2, line = 0.5, las = 3, cex = 0.25);
235 }
236 if (j == 1){
237 mtext(names[i], side = 3, line = 0.5, cex = 0.25);
238 }
239 if (j == length(names)){
240 axis(1, at = (1:short.levels), las = 3, cex.axis = 0.5);
241 }
242 }
243 }
244 colnames(final_pvalue) <- title;
245 #write.table(final_pvalue, file = \"test_final_pvalue.txt\", sep = \"\\t\", quote = FALSE, row.names = FALSE, append = TRUE);
246
247 # get maximum variance larger than expectation by comparison to null bands
248 varnames <- vector();
249 for(i in 1:length(names)){
250 name1 = paste(names[i], \"var\", sep = \"_\")
251 varnames <- c(varnames, name1)
252 }
253 rownames(matrix) <- varnames;
254 colnames(matrix) <- (1:short.levels);
255 max_var <- names;
256 scale <- vector(length = length(names));
257 for (x in 1:nrow(matrix)){
258 if (length(which.max(matrix[x, ])) == 0){
259 scale[x] <- NA;
260 }
261 else{
262 scale[x] <- colnames(matrix)[which.max(matrix[x, ])];
263 }
264 }
265 max_var <- cbind(max_var, scale);
266 write.table(max_var, file = \"$max_dwt\", sep = \"\\t\", quote = FALSE, row.names = FALSE, append = TRUE);
267 return(final_pvalue);
268 }\n";
269
270 print Rcmd "
271 # execute
272 # read in data
273
274 data_test = NULL;
275 data_test <- read.delim(\"$inputFile\");
276
277 pdf(file = \"$pdf\", width = 11, height = 8);
278
279 # loop to read and execute on all $count2 subfiles
280 final = NULL;
281 for (x in 1:$count2){
282 sub = NULL;
283 sub_names = NULL;
284 a = NULL;
285 b = NULL;
286
287 a = ((x - 1) * 12 + 1);
288 b = x * 12;
289
290 if (x < $count2){
291 sub <- data_test[, +c(a:b)];
292 sub_names <- colnames(data_test)[a:b];
293 final <- rbind(final, dwt_var_permut_getMax(sub, sub_names));
294 }
295 else{
296 sub <- data_test[, +c(a:ncol(data_test))];
297 sub_names <- colnames(data_test)[a:ncol(data_test)];
298 final <- rbind(final, dwt_var_permut_getMax(sub, sub_names));
299
300 }
301 }
302
303 dev.off();
304
305 write.table(final, file = \"$pvalue\", sep = \"\\t\", quote = FALSE, row.names = FALSE);
306
307 #eof\n";
308
309 close Rcmd;
310
311 system("echo \"wavelet ANOVA started on \`hostname\` at \`date\`\"\n");
312 system("R --no-restore --no-save --no-readline < $r_script > $r_script.out");
313 system("echo \"wavelet ANOVA ended on \`hostname\` at \`date\`\"\n");
314
315 #close the input and output and error files
316 close(ERROR);
317 close(OUTPUT3);
318 close(OUTPUT2);
319 close(OUTPUT1);
320 close(INPUT);