Mercurial > repos > rhpvorderman > shm_csr

annotate new_imgt.r @ 2:72d68e2e1fd3 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit a8565b413cbe0ef70bc22994ecdd48b400ba6185"
author rhpvorderman
date Fri, 05 Nov 2021 14:43:13 +0000
parents 64d74ba01a7c
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
1 args <- commandArgs(trailingOnly = TRUE)
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
2
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
3 imgt.dir = args[1]
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
4 merged.file = args[2]
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
5 gene = args[3]
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
6
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
7 merged = read.table(merged.file, header=T, sep="\t", fill=T, stringsAsFactors=F, comment.char="", quote="")
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
8
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
9 if(!("Sequence.ID" %in% names(merged))){ #change-o db
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
10 print("Change-O DB changing 'SEQUENCE_ID' to 'Sequence.ID'")
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
11 names(merged)[which(names[merged] == "SEQUENCE_ID")] = "Sequence.ID"
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
12 }
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
13
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
14 if(gene != "-"){
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
15 merged = merged[grepl(paste("^", gene, sep=""), merged$best_match),]
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
16 }
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
17
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
18 if("best_match" %in% names(merged)){
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
19 merged = merged[!grepl("unmatched", merged$best_match),]
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
20 }
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
21
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
22 nrow_dat = 0
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
23
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
24 for(f in list.files(imgt.dir, pattern="*.txt$")){
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
25 #print(paste("filtering", f))
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
26 path = file.path(imgt.dir, f)
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
27 dat = read.table(path, header=T, sep="\t", fill=T, quote="", stringsAsFactors=F, check.names=FALSE, comment.char="")
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
28
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
29 dat = dat[dat[,"Sequence ID"] %in% merged$Sequence.ID,]
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
30
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
31 nrow_dat = nrow(dat)
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
32
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
33 if(nrow(dat) > 0 & grepl("^8_", f)){ #change the FR1 columns to 0 in the "8_..." file
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
34 dat[,grepl("^FR1", names(dat))] = 0
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
35 }
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
36
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
37 write.table(dat, path, quote=F, sep="\t", row.names=F, col.names=T, na="")
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
38 }
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
39
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
40 print(paste("Creating new zip for ", gene, "with", nrow_dat, "sequences"))