|
0
|
1 <tool id="plot_for_lda_output1" name="Draw ROC plot" version="1.0.1">
|
|
|
2 <description>on "Perform LDA" output</description>
|
|
|
3 <command interpreter="sh">r_wrapper.sh $script_file</command>
|
|
|
4
|
|
|
5 <inputs>
|
|
|
6 <param format="txt" name="input" type="data" label="Source file"> </param>
|
|
|
7 <param name="my_title" size="30" type="text" value="My Figure" label="Title of your plot" help="See syntax below"> </param>
|
|
|
8 <param name="X_axis" size="30" type="text" value="Text for X axis" label="Legend of X axis in your plot" help="See syntax below"> </param>
|
|
|
9 <param name="Y_axis" size="30" type="text" value="Text for Y axis" label="Legend of Y axis in your plot" help="See syntax below"> </param>
|
|
|
10 </inputs>
|
|
|
11 <outputs>
|
|
|
12 <data format="pdf" name="pdf_output" />
|
|
|
13 </outputs>
|
|
|
14
|
|
|
15 <tests>
|
|
|
16 <test>
|
|
|
17 <param name="input" value="lda_analy_output.txt"/>
|
|
|
18 <param name="my_title" value="Test Plot1"/>
|
|
|
19 <param name="X_axis" value="Test Plot2"/>
|
|
|
20 <param name="Y_axis" value="Test Plot3"/>
|
|
|
21 <output name="pdf_output" file="plot_for_lda_output.pdf"/>
|
|
|
22 </test>
|
|
|
23 </tests>
|
|
|
24
|
|
|
25 <configfiles>
|
|
|
26 <configfile name="script_file">
|
|
|
27
|
|
|
28 rm(list = objects() )
|
|
|
29
|
|
|
30 ############# FORMAT X DATA #########################
|
|
|
31 format<-function(data) {
|
|
|
32 ind=NULL
|
|
|
33 for(i in 1 : ncol(data)){
|
|
|
34 if (is.na(data[nrow(data),i])) {
|
|
|
35 ind<-c(ind,i)
|
|
|
36 }
|
|
|
37 }
|
|
|
38 #print(is.null(ind))
|
|
|
39 if (!is.null(ind)) {
|
|
|
40 data<-data[,-c(ind)]
|
|
|
41 }
|
|
|
42
|
|
|
43 data
|
|
|
44 }
|
|
|
45
|
|
|
46 ########GET RESPONSES ###############################
|
|
|
47 get_resp<- function(data) {
|
|
|
48 resp1<-as.vector(data[,ncol(data)])
|
|
|
49 resp=numeric(length(resp1))
|
|
|
50 for (i in 1:length(resp1)) {
|
|
|
51 if (resp1[i]=="Control ") {
|
|
|
52 resp[i] = 0
|
|
|
53 }
|
|
|
54 if (resp1[i]=="XLMR ") {
|
|
|
55 resp[i] = 1
|
|
|
56 }
|
|
|
57 }
|
|
|
58 return(resp)
|
|
|
59 }
|
|
|
60
|
|
|
61 ######## CHARS TO NUMBERS ###########################
|
|
|
62 f_to_numbers<- function(F) {
|
|
|
63 ind<-NULL
|
|
|
64 G<-matrix(0,nrow(F), ncol(F))
|
|
|
65 for (i in 1:nrow(F)) {
|
|
|
66 for (j in 1:ncol(F)) {
|
|
|
67 G[i,j]<-as.integer(F[i,j])
|
|
|
68 }
|
|
|
69 }
|
|
|
70 return(G)
|
|
|
71 }
|
|
|
72
|
|
|
73 ###################NORMALIZING#########################
|
|
|
74 norm <- function(M, a=NULL, b=NULL) {
|
|
|
75 C<-NULL
|
|
|
76 ind<-NULL
|
|
|
77
|
|
|
78 for (i in 1: ncol(M)) {
|
|
|
79 if (sd(M[,i])!=0) {
|
|
|
80 M[,i]<-(M[,i]-mean(M[,i]))/sd(M[,i])
|
|
|
81 }
|
|
|
82 # else {print(mean(M[,i]))}
|
|
|
83 }
|
|
|
84 return(M)
|
|
|
85 }
|
|
|
86
|
|
|
87 ##### LDA DIRECTIONS #################################
|
|
|
88 lda_dec <- function(data, k){
|
|
|
89 priors=numeric(k)
|
|
|
90 grandmean<-numeric(ncol(data)-1)
|
|
|
91 means=matrix(0,k,ncol(data)-1)
|
|
|
92 B = matrix(0, ncol(data)-1, ncol(data)-1)
|
|
|
93 N=nrow(data)
|
|
|
94 for (i in 1:k){
|
|
|
95 priors[i]=sum(data[,1]==i)/N
|
|
|
96 grp=subset(data,data\$group==i)
|
|
|
97 means[i,]=mean(grp[,2:ncol(data)])
|
|
|
98 #print(means[i,])
|
|
|
99 #print(priors[i])
|
|
|
100 #print(priors[i]*means[i,])
|
|
|
101 grandmean = priors[i]*means[i,] + grandmean
|
|
|
102 }
|
|
|
103
|
|
|
104 for (i in 1:k) {
|
|
|
105 B= B + priors[i]*((means[i,]-grandmean)%*%t(means[i,]-grandmean))
|
|
|
106 }
|
|
|
107
|
|
|
108 W = var(data[,2:ncol(data)])
|
|
|
109 svdW = svd(W)
|
|
|
110 inv_sqrtW =solve(svdW\$v %*% diag(sqrt(svdW\$d)) %*% t(svdW\$v))
|
|
|
111 B_star= t(inv_sqrtW)%*%B%*%inv_sqrtW
|
|
|
112 B_star_decomp = svd(B_star)
|
|
|
113 directions = inv_sqrtW%*%B_star_decomp\$v
|
|
|
114 return( list(directions, B_star_decomp\$d) )
|
|
|
115 }
|
|
|
116
|
|
|
117 ################ NAIVE BAYES FOR 1D SIR OR LDA ##############
|
|
|
118 naive_bayes_classifier <- function(resp, tr_data, test_data, k=2, tau) {
|
|
|
119 tr_data=data.frame(resp=resp, dir=tr_data)
|
|
|
120 means=numeric(k)
|
|
|
121 #print(k)
|
|
|
122 cl=numeric(k)
|
|
|
123 predclass=numeric(length(test_data))
|
|
|
124 for (i in 1:k) {
|
|
|
125 grp = subset(tr_data, resp==i)
|
|
|
126 means[i] = mean(grp\$dir)
|
|
|
127 #print(i, means[i])
|
|
|
128 }
|
|
|
129 cutoff = tau*means[1]+(1-tau)*means[2]
|
|
|
130 #print(tau)
|
|
|
131 #print(means)
|
|
|
132 #print(cutoff)
|
|
|
133 if (cutoff>means[1]) {
|
|
|
134 cl[1]=1
|
|
|
135 cl[2]=2
|
|
|
136 }
|
|
|
137 else {
|
|
|
138 cl[1]=2
|
|
|
139 cl[2]=1
|
|
|
140 }
|
|
|
141
|
|
|
142 for (i in 1:length(test_data)) {
|
|
|
143
|
|
|
144 if (test_data[i] <= cutoff) {
|
|
|
145 predclass[i] = cl[1]
|
|
|
146 }
|
|
|
147 else {
|
|
|
148 predclass[i] = cl[2]
|
|
|
149 }
|
|
|
150 }
|
|
|
151 #print(means)
|
|
|
152 #print(mean(means))
|
|
|
153 #X11()
|
|
|
154 #plot(test_data,pch=predclass, col=resp)
|
|
|
155 predclass
|
|
|
156 }
|
|
|
157
|
|
|
158 ################# EXTENDED ERROR RATES #################
|
|
|
159 ext_error_rate <- function(predclass, actualclass,msg=c("you forgot the message"), pr=1) {
|
|
|
160 er=sum(predclass != actualclass)/length(predclass)
|
|
|
161
|
|
|
162 matr<-data.frame(predclass=predclass,actualclass=actualclass)
|
|
|
163 escapes = subset(matr, actualclass==1)
|
|
|
164 subjects = subset(matr, actualclass==2)
|
|
|
165 er_esc=sum(escapes\$predclass != escapes\$actualclass)/length(escapes\$predclass)
|
|
|
166 er_subj=sum(subjects\$predclass != subjects\$actualclass)/length(subjects\$predclass)
|
|
|
167
|
|
|
168 if (pr==1) {
|
|
|
169 # print(paste(c(msg, 'overall : ', (1-er)*100, "%."),collapse=" "))
|
|
|
170 # print(paste(c(msg, 'within escapes : ', (1-er_esc)*100, "%."),collapse=" "))
|
|
|
171 # print(paste(c(msg, 'within subjects: ', (1-er_subj)*100, "%."),collapse=" "))
|
|
|
172 }
|
|
|
173 return(c((1-er)*100, (1-er_esc)*100, (1-er_subj)*100))
|
|
|
174 }
|
|
|
175
|
|
|
176 ## Main Function ##
|
|
|
177
|
|
|
178 files_alias<-c("${my_title}")
|
|
|
179 tau=seq(0,1,by=0.005)
|
|
|
180 nfiles=1
|
|
|
181 f = c("${input}")
|
|
|
182
|
|
|
183 rez_ext<-list()
|
|
|
184 for (i in 1:nfiles) {
|
|
|
185 rez_ext[[i]]<-dget(paste(f[i], sep="",collapse=""))
|
|
|
186 }
|
|
|
187
|
|
|
188 tau<-tau[1:(length(tau)-1)]
|
|
|
189 for (i in 1:nfiles) {
|
|
|
190 rez_ext[[i]]<-rez_ext[[i]][,1:(length(tau)-1)]
|
|
|
191 }
|
|
|
192
|
|
|
193 ######## OPTIMAIL TAU ###########################
|
|
|
194
|
|
|
195 #rez_ext
|
|
|
196
|
|
|
197 rate<-c("Optimal tau","Tr total", "Tr Y", "Tr X")
|
|
|
198
|
|
|
199 m_tr<-numeric(nfiles)
|
|
|
200 m_xp22<-numeric(nfiles)
|
|
|
201 m_x<-numeric(nfiles)
|
|
|
202
|
|
|
203 for (i in 1:nfiles) {
|
|
|
204 r<-rez_ext[[i]]
|
|
|
205 #tr
|
|
|
206 # rate<-rbind(rate, c(files_alias[i]," "," "," ") )
|
|
|
207 mm<-which((r[3,])==max(r[3,]))
|
|
|
208
|
|
|
209 m_tr[i]<-mm[1]
|
|
|
210 rate<-rbind(rate,c(tau[m_tr[i]],r[,m_tr[i]]))
|
|
|
211 }
|
|
|
212 print(rate)
|
|
|
213
|
|
|
214 pdf(file= paste("${pdf_output}"))
|
|
|
215
|
|
|
216 plot(rez_ext[[i]][2,]~rez_ext[[i]][3,], xlim=c(0,100), ylim=c(0,100), xlab="${X_axis} [1-FP(False Positive)]", ylab="${Y_axis} [1-FP(False Positive)]", type="l", lty=1, col="blue", xaxt='n', yaxt='n')
|
|
|
217 for (i in 1:nfiles) {
|
|
|
218 lines(rez_ext[[i]][2,]~rez_ext[[i]][3,], xlab="${X_axis} [1-FP(False Positive)]", ylab="${Y_axis} [1-FP(False Positive)]", type="l", lty=1, col=i)
|
|
|
219 # pt=c(r,)
|
|
|
220 points(x=rez_ext[[i]][3,m_tr[i]],y=rez_ext[[i]][2,m_tr[i]], pch=16, col=i)
|
|
|
221 }
|
|
|
222
|
|
|
223
|
|
|
224 title(main="${my_title}", adj=0, cex.main=1.1)
|
|
|
225 axis(2, at=c(0,20,40,60,80,100), labels=c('0','20','40','60','80','100%'))
|
|
|
226 axis(1, at=c(0,20,40,60,80,100), labels=c('0','20','40','60','80','100%'))
|
|
|
227
|
|
|
228 #leg=c("10 kb","50 kb","100 kb")
|
|
|
229 #legend("bottomleft",legend=leg , col=c(1,2,3), lty=c(1,1,1))
|
|
|
230
|
|
|
231 #dev.off()
|
|
|
232
|
|
|
233 </configfile>
|
|
|
234 </configfiles>
|
|
|
235
|
|
|
236
|
|
|
237 <help>
|
|
|
238 .. class:: infomark
|
|
|
239
|
|
|
240 **What it does**
|
|
|
241
|
|
|
242 This tool generates a Receiver Operating Characteristic (ROC) plot that shows LDA classification success rates for different values of the tuning parameter tau as Figure 3 in Carrel et al., 2006 (PMID: 17009873).
|
|
|
243
|
|
|
244 *Carrel L, Park C, Tyekucheva S, Dunn J, Chiaromonte F, et al. (2006) Genomic Environment Predicts Expression Patterns on the Human Inactive X Chromosome. PLoS Genet 2(9): e151. doi:10.1371/journal.pgen.0020151*
|
|
|
245
|
|
|
246 -----
|
|
|
247
|
|
|
248 .. class:: warningmark
|
|
|
249
|
|
|
250 **Note**
|
|
|
251
|
|
|
252 - Output from "Perform LDA" tool is used as input file for this tool.
|
|
|
253
|
|
|
254 </help>
|
|
|
255
|
|
|
256
|
|
|
257
|
|
|
258 </tool>
|
