diff all stuff/Difference finder for GalaxyP 2-21-20.R @ 11:77e47268b650 draft default tip

Uploaded
author jfb
date Tue, 14 Jul 2020 19:53:34 -0400
parents 78f16bc92ba6
children
line wrap: on
line diff
--- a/all stuff/Difference finder for GalaxyP 2-21-20.R	Fri Feb 21 16:37:08 2020 -0500
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,358 +0,0 @@
-#ff you want ONLY FULL MOTIFS, put "YES" here, please use all caps
-FullMotifsOnly_questionmark<-"NO"
-#If you want ONLY TRUNCATED MOTIFS, put "YES" here, please use all caps
-TruncatedMotifsOnly_questionmark<-"NO"
-#if you want to find the overlap, put a "YES" here (all caps), if you want to find the non-overlap, put "NO" (all caps)
-Are_You_Looking_For_Commonality<-"NO"
-
-
-#put the names of your input files here
-FirstSubstrateSet<- read.csv("S1.csv", stringsAsFactors=FALSE,colClasses = "character")
-Firstsubbackfreq<- read.csv("SBF1.csv", header=FALSE, stringsAsFactors=FALSE)
-
-SecondSubstrateSet<- read.csv("S2.csv", stringsAsFactors=FALSE,colClasses = "character")
-Secondsubbackfreq<- read.csv("SBF2.csv", header=FALSE, stringsAsFactors=FALSE)
-
-
-First_unshared_motifs_table<-"1RS.csv"
-First_unshared_subbackfreq<-"1RSBF.csv"
-
-Second_unshared_motifs_table<-"2RS.csv"
-Second_unshared_subbackfreq<-"2RSBF.csv"
-
-EmptySubHeader<-colnames(FirstSubstrateSet)
-EmptySubHeader<-matrix(EmptySubHeader, nrow=1)
-EmptySBFHeader<-Firstsubbackfreq[,1]
-
-#final note, this code is going to be unworkable if you want to make a Venn diagram of more than 3 circles.  I think I'll poke around
-#other languages to see if any of them can do it.
-####################################################################################################################################
-
-
-FirstCentralLetters<-FirstSubstrateSet[,11]
-SecondCentralLetters<-SecondSubstrateSet[,11]
-
-FirstEsses<-sapply(FirstCentralLetters, grepl, pattern="S", ignore.case=TRUE)
-FirstTees<-sapply(FirstCentralLetters, grepl, pattern="T", ignore.case=TRUE)
-FirstWys<-sapply(FirstCentralLetters, grepl, pattern="Y", ignore.case=TRUE)
-
-SecondEsses<-sapply(SecondCentralLetters, grepl, pattern="S", ignore.case=TRUE)
-SecondTees<-sapply(SecondCentralLetters, grepl, pattern="T", ignore.case=TRUE)
-SecondWys<-sapply(SecondCentralLetters, grepl, pattern="Y", ignore.case=TRUE)
-
-FirstCentralLetters<-replace(FirstCentralLetters,FirstEsses,"xS")
-FirstCentralLetters<-replace(FirstCentralLetters,FirstTees,"xT")
-FirstCentralLetters<-replace(FirstCentralLetters,FirstWys,"xY")
-
-SecondCentralLetters<-replace(SecondCentralLetters,SecondEsses,"xS")
-SecondCentralLetters<-replace(SecondCentralLetters,SecondTees,"xT")
-SecondCentralLetters<-replace(SecondCentralLetters,SecondWys,"xY")
-
-FirstCentralLetters->FirstSubstrateSet[,11]
-SecondCentralLetters->SecondSubstrateSet[,11]
-
-####################################################################################################################################
-####################################################################################################################################
-# better version of this code written in C: what happens when two kinases share a motif, but they found that motif in two 
-# separate proteins thus two separate accession numbers?
-# It should actually output the shared motif and BOTH accession numbers.  Right now it does not, it only maps out the second
-# accession number.  So that needs to be fixed BUT you need to keep the commonality between a motif and its accession number
-####################################################################################################################################
-####################################################################################################################################
-####################################################################################################################################
-####################################################################################################################################
-
-#Create the motif sets, deciding wether or not you're looking for truncated or full here
-#full only
-
-
-###############################################
-#ALL motifs, full and truncated
-
-if (FullMotifsOnly_questionmark!="YES"&&TruncatedMotifsOnly_questionmark!="YES"){
-  FTLwtmotifs=matrix(,nrow = nrow(FirstSubstrateSet),ncol=1)
-  FTLwtAccessionNumbers=matrix(,nrow = nrow(FirstSubstrateSet),ncol=1)
-  
-  for (i in 1:nrow(FirstSubstrateSet)){
-    FTLwtletters<-FirstSubstrateSet[i,4:18]
-    FTLwtletters<-FTLwtletters[FTLwtletters !="XXXXX"]
-    FTLwtletters<-paste(FTLwtletters, sep="", collapse="")
-    leftspaces<-c()
-    rightspaces<-c()
-    
-    YYYmotif <- unlist(strsplit(FTLwtletters, split = ""))
-    YYYposition <- match(x = "x", table = YYYmotif)
-    #position itself tells me how much is to the left of that X by what it's number is.  x at position 4 tells me that there are
-    #just 3 letters to the left of x
-    
-    YYYLettersToTheLeft <- YYYposition - 1
-    #how many letters to the right SHOULD just be length(motif)-position-1 if it's 5 long and x is at 3 then Y is at 4 and there is
-    #just 1 spot to the right of Y so LettersToTheRight<-1 because 5-3-1=1
-    YYYLettersToTheRight <- length(YYYmotif) - YYYposition - 1
-    #then sanity check, we're currently looking only at +/-4, but this spot allows for up to +/- 7 as well, just depends on what the
-    #variable the user puts in is
-    
-    
-    if (YYYLettersToTheLeft < 7 | YYYLettersToTheRight < 7) {
-      leftspaces<-rep(" ",times=(7-YYYLettersToTheLeft))
-      rightspaces<-rep(" ",times=7-(YYYLettersToTheRight))
-      #add blank spaces if the motif has less than 4 letters to the left/right
-      motif<-c(leftspaces,YYYmotif,rightspaces)
-      #save that motif, which is the Y and +/- 4 amino acids, including truncation
-      motif<-motif[!motif %in% "x"]
-      motif<-paste(motif, sep="", collapse="")
-      FTLwtletters<-motif
-      FTLwtmotifs[i,1]<-FTLwtletters
-      FTLwtAccessionNumbers[i,1]<-FirstSubstrateSet[i,3]
-    }
-    
-    if(YYYLettersToTheLeft>6 && YYYLettersToTheRight>6){
-      motif<-YYYmotif
-      #add blank spaces if the motif has less than 4 letters to the left/right
-      motif<-c(leftspaces,YYYmotif,rightspaces)
-      #save that motif, which is the Y and +/- 4 amino acids, including truncation
-      motif<-motif[!motif %in% "x"]
-      motif<-paste(motif, sep="", collapse="")
-      FTLwtletters<-motif
-      FTLwtmotifs[i,1]<-FTLwtletters
-      FTLwtAccessionNumbers[i,1]<-FirstSubstrateSet[i,3]
-      
-      
-    }
-    
-  }
-  
-  D835Ymotifs=matrix(,nrow = nrow(SecondSubstrateSet),ncol=1)
-  D835YAccessionNumbers<-matrix(,nrow = nrow(SecondSubstrateSet),ncol = 1)
-  
-  for (i in 1:nrow(SecondSubstrateSet)){
-    D835letters<-SecondSubstrateSet[i,4:18]
-    D835letters<-D835letters[D835letters !="XXXXX"]
-    D835letters<-paste(D835letters, sep="", collapse="")
-    leftspaces<-c()
-    rightspaces<-c()
-    
-    YYYmotif <- unlist(strsplit(D835letters, split = ""))
-    YYYposition <- match(x = "x", table = YYYmotif)
-    #position itself tells me how much is to the left of that X by what it's number is.  x at position 4 tells me that there are
-    #just 3 letters to the left of x
-    
-    YYYLettersToTheLeft <- YYYposition - 1
-    #how many letters to the right SHOULD just be length(motif)-position-1 if it's 5 long and x is at 3 then Y is at 4 and there is
-    #just 1 spot to the right of Y so LettersToTheRight<-1 because 5-3-1=1
-    YYYLettersToTheRight <- length(YYYmotif) - YYYposition - 1
-    #then sanity check, we're currently looking only at +/-4, but this spot allows for up to +/- 7 as well, just depends on what the
-    #variable the user puts in is
-    if (YYYLettersToTheLeft < 7 | YYYLettersToTheRight < 7) {
-      leftspaces<-rep(" ",times=(7-YYYLettersToTheLeft))
-      rightspaces<-rep(" ",times=7-(YYYLettersToTheRight))
-      #add blank spaces if the motif has less than 4 letters to the left/right
-      motif<-c(leftspaces,YYYmotif,rightspaces)
-      #save that motif, which is the Y and +/- 4 amino acids, including truncation
-      motif<-motif[!motif %in% "x"]
-      motif<-paste(motif, sep="", collapse="")
-      D835letters<-motif
-      D835Ymotifs[i,1]<-D835letters
-      D835YAccessionNumbers[i,1]<-SecondSubstrateSet[i,3]
-    }
-    
-    if(YYYLettersToTheLeft>6 && YYYLettersToTheRight>6){
-      motif<-YYYmotif
-      #add blank spaces if the motif has less than 4 letters to the left/right
-      motif<-c(leftspaces,YYYmotif,rightspaces)
-      #save that motif, which is the Y and +/- 4 amino acids, including truncation
-      motif<-motif[!motif %in% "x"]
-      motif<-paste(motif, sep="", collapse="")
-      D835letters<-motif
-      D835Ymotifs[i,1]<-D835letters
-      D835YAccessionNumbers[i,1]<-SecondSubstrateSet[i,3]
-    }
-  }
-  
-  names(FTLwtmotifs)<-FTLwtAccessionNumbers
-  names(D835Ymotifs)<-D835YAccessionNumbers
-  
-}
-
-
-FTLwtmotifsFINAL<-FTLwtmotifs[!FTLwtmotifs %in% D835Ymotifs]
-FTLwtmotifsFINAL<-FTLwtmotifsFINAL[!duplicated(FTLwtmotifsFINAL)]
-
-D835YmotifsFINAL<-D835Ymotifs[!D835Ymotifs %in% FTLwtmotifs]
-D835YmotifsFINAL<-D835YmotifsFINAL[!duplicated(D835YmotifsFINAL)]
-
-
-columnalheader<-c(rep(NA,36))
-FTLFinalMatrix<-matrix(data =columnalheader,nrow = 1)
-
-FLTheader<-c("Substrate","Species","Reference","-7","-6","-5","-4","-3","-2","-1","0","1","2","3","4","5","6","7","Phosphite")
-
-if (length(FTLwtmotifsFINAL)>0){
-  for (k in 1:length(FTLwtmotifsFINAL)) {
-    AN<-00000
-    #I don't remember why, but I felt it necessary to destroy the accession number multiple times to ensure it is
-    #destroyed immediately after use
-    for (m in 1:ncol(Firstsubbackfreq)) {
-      AN <- as.character(Firstsubbackfreq[1, m])
-      if (grepl(pattern = AN,
-                x = names(FTLwtmotifsFINAL[k]),
-                fixed = TRUE) == TRUE) {
-        outputmatrix <- as.character(Firstsubbackfreq[, m])
-        outputmatrix <- matrix(outputmatrix, nrow = 1)
-        #with that accession number, find a match in the subbackfreq file and save it here
-        FTLFinalMatrix<-rbind(FTLFinalMatrix,outputmatrix)
-      }
-    }
-  }
-  FTLFinalMatrix<-FTLFinalMatrix[!duplicated(FTLFinalMatrix),]
-  FTLFinalMatrix<-FTLFinalMatrix[2:nrow(FTLFinalMatrix),]
-  
-  
-  FTLoutputmatrix<-matrix(data=c(FTLwtmotifsFINAL,names(FTLwtmotifsFINAL)),ncol = 2)
-  # FLTheader<-unlist(FLTheader)
-  lefthandFLT<-matrix(data = rep(NA,times=2*nrow(FTLoutputmatrix)),nrow=nrow(FTLoutputmatrix))
-  righthandFLT<-matrix(data = rep(NA,times=1*nrow(FTLoutputmatrix)),nrow=nrow(FTLoutputmatrix))
-  FLTaccessionset<-FTLoutputmatrix[,2]
-  FTLmeat<-sapply(FTLoutputmatrix[,1], strsplit, "")
-  FTLmeat<-sapply(FTLmeat, unlist)
-  colnames(FTLmeat)<-NULL
-  FTLmeat<-t(FTLmeat)
-  
-  FTLoutputmatrix2<-cbind(lefthandFLT,FLTaccessionset,FTLmeat,righthandFLT)
-  colnames(FTLoutputmatrix2)<-NULL
-  rownames(FTLoutputmatrix2)<-NULL
-  colnames(FLTheader)<-NULL
-  rownames(FLTheader)<-NULL
-  
-  
-  FirstCentralLettersAGAIN<-FTLoutputmatrix2[,11]
-  
-  FirstEsses<-sapply(FirstCentralLettersAGAIN, grepl, pattern="S", ignore.case=TRUE)
-  FirstTees<-sapply(FirstCentralLettersAGAIN, grepl, pattern="T", ignore.case=TRUE)
-  FirstWys<-sapply(FirstCentralLettersAGAIN, grepl, pattern="Y", ignore.case=TRUE)
-  
-  FirstCentralLettersAGAIN<-replace(FirstCentralLettersAGAIN,FirstEsses,"xS")
-  FirstCentralLettersAGAIN<-replace(FirstCentralLettersAGAIN,FirstTees,"xT")
-  FirstCentralLettersAGAIN<-replace(FirstCentralLettersAGAIN,FirstWys,"xY")
-  
-  FirstCentralLettersAGAIN->FTLoutputmatrix2[,11]
-  
-  FTLoutputmatrix2<-rbind(FLTheader,FTLoutputmatrix2)
-  
-  write.table(x=FTLoutputmatrix2,
-              file=First_unshared_motifs_table,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-  
-  columnalheader<-c(as.character(Firstsubbackfreq[1:36,1]))
-  columnalheader<-matrix(columnalheader,nrow = 1)
-  write.table(x=columnalheader,
-              file=First_unshared_subbackfreq,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-  
-  write.table(x=FTLFinalMatrix,
-              file=First_unshared_subbackfreq,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-} else{
-  FTLFinalMatrix<-columnalheader
-  write.table(x=EmptySubHeader,
-              file=First_unshared_motifs_table,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-  
-  columnalheader<-c(as.character(Firstsubbackfreq[1:36,1]))
-  columnalheader<-matrix(columnalheader,nrow = 1)
-  write.table(x=columnalheader,
-              file=First_unshared_subbackfreq,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-}
-
-
-columnalheader<-c(rep(NA,36))
-D835YFinalMatrix<-matrix(data =columnalheader,nrow = 1)
-
-if (length(D835YmotifsFINAL)>0){
-  for (k in 1:length(D835YmotifsFINAL)) {
-    #I don't remember why, but I felt it necessary to destroy the accession number multiple times to ensure it is
-    #destroyed immediately after use
-    for (m in 1:ncol(Secondsubbackfreq)) {
-      AN <- as.character(Secondsubbackfreq[1, m])
-      if (grepl(pattern = AN,
-                x = names(D835YmotifsFINAL[k]),
-                fixed = TRUE) == TRUE) {
-        outputmatrix <- as.character(Secondsubbackfreq[, m])
-        outputmatrix <- matrix(outputmatrix, nrow = 1)
-        #with that accession number, find a match in the subbackfreq file and save it here
-        D835YFinalMatrix<-rbind(D835YFinalMatrix,outputmatrix)
-      }
-    }
-  }
-  D835YFinalMatrix<-D835YFinalMatrix[!duplicated(D835YFinalMatrix),]
-  D835YFinalMatrix<-D835YFinalMatrix[2:nrow(D835YFinalMatrix),]
-  
-  D835Youtputmatrix<-matrix(data=c(D835YmotifsFINAL,names(D835YmotifsFINAL)),ncol = 2)
-  
-  D835Yheader<-c("Substrate","Species","Reference","-7","-6","-5","-4","-3","-2","-1","0","1","2","3","4","5","6","7","Phosphite")
-  # D835Yheader<-unlist(D835Yheader)
-  lefthandD835<-matrix(data = rep(NA,times=2*nrow(D835Youtputmatrix)),nrow=nrow(D835Youtputmatrix))
-  righthandD835<-matrix(data = rep(NA,times=1*nrow(D835Youtputmatrix)),nrow=nrow(D835Youtputmatrix))
-  D835Yaset<-D835Youtputmatrix[,2]
-  D835meat<-sapply(D835Youtputmatrix[,1], strsplit, "")
-  D835meat<-sapply(D835meat, unlist)
-  colnames(D835meat)<-NULL
-  D835meat<-t(D835meat)
-  
-  D835Youtputmatrix2<-cbind(lefthandD835,D835Yaset,D835meat,righthandD835)
-  colnames(D835Youtputmatrix2)<-NULL
-  rownames(D835Youtputmatrix2)<-NULL
-  colnames(D835Yheader)<-NULL
-  rownames(D835Yheader)<-NULL
-  
-  
-  SecondCentralLettersAGAIN<-D835Youtputmatrix2[,11]
-  
-  SecondEsses<-sapply(SecondCentralLettersAGAIN, grepl, pattern="S", ignore.case=TRUE)
-  SecondTees<-sapply(SecondCentralLettersAGAIN, grepl, pattern="T", ignore.case=TRUE)
-  SecondWys<-sapply(SecondCentralLettersAGAIN, grepl, pattern="Y", ignore.case=TRUE)
-  
-  SecondCentralLettersAGAIN<-replace(SecondCentralLettersAGAIN,SecondEsses,"xS")
-  SecondCentralLettersAGAIN<-replace(SecondCentralLettersAGAIN,SecondTees,"xT")
-  SecondCentralLettersAGAIN<-replace(SecondCentralLettersAGAIN,SecondWys,"xY")
-  
-  SecondCentralLettersAGAIN->D835Youtputmatrix2[,11]
-  
-  D835Youtputmatrix2<-rbind(D835Yheader,D835Youtputmatrix2)
-  
-  write.table(x=D835Youtputmatrix2,
-              file=Second_unshared_motifs_table,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-  
-  columnalheader<-c(as.character(Firstsubbackfreq[1:36,1]))
-  columnalheader<-matrix(columnalheader,nrow = 1)
-  write.table(x=columnalheader,
-              file=Second_unshared_subbackfreq,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-  
-  write.table(x=D835YFinalMatrix,
-              file=Second_unshared_subbackfreq,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-} else {
-  D835YFinalMatrix<- columnalheader
-  write.table(x=EmptySubHeader,
-              file=Second_unshared_motifs_table,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-  
-  columnalheader<-c(as.character(Firstsubbackfreq[1:36,1]))
-  columnalheader<-matrix(columnalheader,nrow = 1)
-  write.table(x=columnalheader,
-              file=Second_unshared_subbackfreq,
-              quote=FALSE, sep=",",
-              row.names=FALSE,col.names = FALSE, na="", append=TRUE)
-}
\ No newline at end of file