Efrain Gonzalez / Cleaning and Fixing Data with R

Download as
Browse Code ยป

Commit 58ba5d0b06bb84f11c3fc28a88967881141849b5

Authored by Efrain Gonzalez
1 parent 18c7602e69
Exists in master

New version which includes the storing of the clean version of the GPL file (UNTESTED)

Showing 1 changed file with 302 additions and 0 deletions   Show diff stats
File was created 1 #Libraries required to run the code
2 library(pryr)
3 library(MASS)
4 library(dplyr)
5 library(tidyr)
6 library(readr)
7 library(stringr)
8
9
10 #Necessary Functions
11 #1#Function for handling the changing of row names and column names
12 chngrownm <- function(mat){
13 row <- dim(mat)[1]
14 col <- dim(mat)[2]
15 j <- 1
16 x <- 1
17 p <- 1
18 a <- 1
19 b <- 1
20 g <- 1
21 for(j in 1:col){
22 if("!Sample_source_name_ch1"==mat[1,j]){
23 colnames(mat)[j] <- "Brain_Region"
24 }
25 if("!Sample_title" == mat[1,j]){
26 colnames(mat)[j] <- "Title"
27 }
28 if("!Sample_geo_accession" == mat[1,j]){
29 colnames(mat)[j] <- "ID_REF"
30 } else{
31 if(grepl("Sex|gender|Gender|sex",mat[2,j])==TRUE){
32 colnames(mat)[j] <- paste0("Sex",x)
33 x = x + 1
34 }
35 if(grepl("postmorteminterval|PMI|pmi",mat[2,j])==TRUE){
36 colnames(mat)[j] <- paste0("PMI",p)
37 p = p + 1
38 }
39 if(grepl("age|Age|AGE",mat[2,j])==TRUE){
40 colnames(mat)[j] <- paste0("Age",a)
41 a = a + 1
42 }
43 if(grepl("braak|b&b",mat[2,j])==TRUE){
44 colnames(mat)[j] <- paste0("Braak",b)
45 b = b + 1
46 }
47 if(grepl("group|disease|control|AD|normal|diagnosis|Alzheimer|Control",mat[2,j])==TRUE){
48 colnames(mat)[j] <- paste0("Group",g)
49 g = g + 1
50 }
51
52 }
53 j = j + 1
54 }
55 mat
56 }
57
58 #2#Function for reorganizing information within the columns
59 cinfo <- function(mat){
60 col <- dim(mat)[2]
61 j <-2
62 for(j in 2:col){
63 if(grepl("Group",colnames(mat)[j]) == TRUE){
64 mat[,j] <- gsub(".+:\\s|\\s.+;.+","",mat[,j])
65 }
66 if(grepl("Age",colnames(mat)[j])==TRUE){
67 mat[,j] <- gsub("\\D","",mat[,j])%>%
68 as.integer()
69 }
70 if(grepl("Sex",colnames(mat)[j])==TRUE){
71 mat[,j] <- gsub(".+:\\s","",mat[,j])
72 }
73 if(grepl("PMI",colnames(mat)[j])==TRUE){
74 mat[,j] <- gsub("[^0-9\\.]","",mat[,j])%>%
75 as.numeric()
76 }
77 if(grepl("Braak",colnames(mat)[j])==TRUE){
78 mat[,j]<-gsub(".+:\\s","",mat[,j])%>%
79 as.roman()%>%
80 as.integer()
81 }
82 j=j+1
83 }
84 mat
85 }
86
87 #3#Function for labeling the gene IDs without names
88 NAFIXING <- function(GIDNAM){
89 row <- dim(GIDNAM)[1]
90 i <- 1
91 for(i in 1:row){
92 if(grepl("^NA\\s*$",GIDNAM[i,2])==TRUE||is.na(GIDNAM[i,2])==TRUE){
93 GIDNAM[i,2] <- GIDNAM[i,1]
94 }
95 i <- i + 1
96 }
97 GIDNAM
98 }
99
100 #4#Function for changing the gene ID to gene name
101 cgeneID <- function(GeneName,DATA){
102 colGene <- dim(GeneName)[2]
103 j <- 1
104 for(j in 1:colGene){
105 chngsreq <- grep(GeneName[1,j],DATA[1,])
106 #DATA[1,chngsreq] <- gsub(GeneName[1,j],GeneName[2,j],DATA[1,chngsreq])
107 DATA[1,chngsreq] <- gsub(paste0("^",GeneName[1,j]),GeneName[2,j],DATA[1,chngsreq])
108 j = j+1
109 }
110 DATA
111 }
112
113 #5#Function for adjusting the gene names
114 gcnames <- function(DiData,usecol=1){
115 nuruns <- dim(DiData)[2]
116 i = 1
117 nwnam <- rep("0",length.out=nuruns)
118 for(i in 1:nuruns){
119 if(length(strsplit(colnames(DiData)[i],"///")[[1]]) >= usecol){
120 nwnam[i]=strsplit(colnames(DiData)[i],"///")[[1]][usecol]
121 } else{
122 nwnam[i]=strsplit(colnames(DiData)[i],"///")[[1]][1]
123 }
124
125 }
126 nwnam
127
128 }
129
130
131
132 #The Rest of this code will be used every time you want to change a data set
133
134 #Getting the series matrix file
135 print("Choose the series matrix file that you want to Analyze")
136 alz <- file.choose()
137
138 #Getting the GPL file
139 print("Choose the GPL file that correlates with the above series matrix file")
140 genena <- file.choose()
141
142
143 #Set working directory based on the directory of the series matrix file Currently only works for windows
144 ##strsplit(alz,"[\\]") %>%
145 ## .[[1]] %>%
146 ## .[-length(.)] %>%
147 ## paste(.,collapse="/") %>%
148 ## setwd()
149
150 #Find out if it is a soft GPL file or not
151 soft <- strsplit(genena,"[\\|/]") %>%
152 .[[1]] %>%
153 .[length(.)] %>%
154 grepl("soft",.)
155
156 #Working with the wordy part of the document
157 alzword <- alz %>%
158 read_delim(delim ="\t",comment = "!Series",col_names = FALSE)%>%
159 filter(grepl("!Sample",X1))%>%
160 filter(!grepl("!Sample_contact",X1))
161
162 ##Changing row names and column names:
163 ALZWORD <- t(alzword)
164 rownames(ALZWORD)=NULL
165 colnames(ALZWORD) <- colnames(ALZWORD,do.NULL=FALSE)
166 ALZWORD <- chngrownm(ALZWORD)[-1,]
167 ALZWORD <- ALZWORD%>%
168 as.data.frame()%>%
169 dplyr::select(-starts_with("col"))
170
171 ##Reorganizing information within the columns
172 ALZWORDF <- cinfo(ALZWORD)
173
174
175 #Working with Actual Data part of file
176 alzdat <- alz %>%
177 read_delim(delim="\t",col_names=TRUE,comment = "!",skip=1)
178 ALZDAT <- t(alzdat[,-1])
179 rownames(ALZDAT)=NULL
180
181 ##Is there a clean version of the GPL file available?
182 gplnum <- strsplit(genena,"[\\|/]") %>%
183 .[[1]] %>%
184 .[length(.)] %>%
185 gsub("\\D","",.)
186 clfileex <- sum(grepl(paste0("Clean_GPL",gplnum),list.files()))
187 if(clfileex >= 1){
188 #use the clean version
189 geneIDNam <- paste0("Clean_GPL",gplnum,".txt") %>%
190 read_delim(delim="\t",col_names = c("ID","Symbol"), comment = "!")
191
192 }
193 if(clfileex == 0){
194 ##Lets Create a clean version
195
196 ##Gene ID to Gene Name
197 ###geneIDNam <- genena %>%
198 ### read_delim(delim="\t",comment = "#")%>%
199 ### dplyr::select(.,ID,grep("Symbol|ORF",colnames(.)))
200 ###problems with the above for soft files
201 if(soft == TRUE){
202 #gplnum <- strsplit(genena,"[\\|/]") %>%
203 # .[[1]] %>%
204 # .[length(.)] %>%
205 # gsub("\\D","",.)
206 #Check to see if there is already a file containing information on soft files
207 fileex <- sum(grepl("GPL_ID_LOC.txt",list.files()))
208 if(fileex == 1){
209 #Check to see if this GPL soft file has been used before
210 IDF <- read_delim("GPL_ID_LOC.txt",delim = "\t",col_names = TRUE) %>%
211 .$GPL_FILE_NUM%>%
212 grepl(gplnum,.) %>%
213 sum()
214 if(IDF == 1){
215 IDLOCAL <- read_delim("GPL_ID_LOC.txt",delim = "\t",col_names = TRUE) %>%
216 .$GPL_FILE_NUM%>%
217 grep(gplnum,.)
218 idlocgpl <- read_delim("GPL_ID_LOC.txt",delim = "\t",col_names = TRUE) %>%
219 .$LOC_ID %>%
220 .[IDLOCAL]
221 geneIDNam <- genena %>%
222 read_delim(delim="\t",col_names = TRUE, comment = "!", skip = idlocgpl) %>%
223 dplyr::select(.,ID,grep("Symbol|ORF",colnames(.)))
224 }
225 if(IDF == 0){
226 #No information on this particular GPL file
227 idLOCGPL <- genena %>%
228 read_delim(delim="\t",col_names = FALSE, comment = "!", n_max = 1000) %>%
229 t(.) %>%
230 grep("^\\D",.) %>%
231 length()-1
232 cbind(as.integer(gplnum),as.integer(idLOCGPL)) %>%
233 cat(file="GPL_ID_LOC.txt",sep = "\t", fill = TRUE, append = TRUE)
234 geneIDNam <- genena %>%
235 read_delim(delim="\t",col_names = TRUE, comment = "!", skip = idLOCGPL) %>%
236 dplyr::select(.,ID,grep("Symbol|ORF",colnames(.)))
237 }
238 }
239 if(fileex == 0){
240 #We must create a file that we can access for later use
241 idLOCGPL <- genena %>%
242 read_delim(delim="\t",col_names = FALSE, comment = "!", n_max = 1000) %>%
243 t(.) %>%
244 grep("^\\D",.) %>%
245 length()-1
246 Firstval <- cbind(as.integer(gplnum),as.integer(idLOCGPL))
247 colnames(Firstval) <- c("GPL_FILE_NUM","LOC_ID")
248 write.table(Firstval,file = "GPL_ID_LOC.txt", sep = "\t",row.names = FALSE, col.names = TRUE)
249 geneIDNam <- genena %>%
250 read_delim(delim="\t",col_names = TRUE, comment = "!", skip = idLOCGPL) %>%
251 dplyr::select(.,ID,grep("Symbol|ORF",colnames(.)))
252 }
253 }
254 if(soft == FALSE){
255 geneIDNam <- genena %>%
256 read_delim(delim="\t",comment = "#")%>%
257 dplyr::select(.,ID,grep("Symbol|ORF",colnames(.)))
258 }
259
260 ##Labeling the gene IDs without names
261 geneIDNam <- NAFIXING(geneIDNam)
262
263 ##remove the whitespace
264 geneIDNam <- t(rbind(str_trim(t(geneIDNam)[1,]),str_trim(t(geneIDNam)[2,])))
265
266 ##Here is the clean version
267 write.table(geneIDNam,file = paste0("Clean_GPL",gplnum,".txt"),sep = "\t",row.names = FALSE, col.names = FALSE)
268 }
269
270
271
272 ##Changing the gene ID to gene name
273 ALZDAT1 <- cgeneID(t(geneIDNam),t(alzdat))
274 colnames(ALZDAT) = ALZDAT1[1,]
275
276
277 ##Adjusting the column names aka the gene names
278 colnames(ALZDAT) <- gcnames(ALZDAT)
279
280
281 #Full Data
282 Fullalzdw <- ALZDAT %>%
283 as.data.frame() %>%
284 cbind(ALZWORDF,.)
285
286
287 #nfna <- strsplit(alz,"[\\|/]") %>%
288 # .[[1]] %>%
289 # .[length(.)] %>%
290 # gsub("\\D","",.) %>%
291 # c("GSE",.,"after.txt") %>%
292 # paste(collapse = "")
293 #write.matrix(Fullalzdw,file = nfna,sep = "\t")
294
295 #Perfect for excel viewing
296 nfnaex <- strsplit(alz,"[\\]") %>%
297 .[[1]] %>%
298 .[length(.)] %>%
299 gsub("\\D","",.) %>%
300 c("GSE",.,"aftexcel.txt") %>%
301 paste(collapse = "")
302 write.table(t(Fullalzdw), file = nfnaex, sep = "\t",row.names = FALSE)