|
0
|
1 args <- commandArgs(trailingOnly = TRUE)
|
|
|
2
|
|
|
3
|
|
|
4 summaryfile = args[1]
|
|
|
5 sequencesfile = args[2]
|
|
|
6 mutationanalysisfile = args[3]
|
|
|
7 mutationstatsfile = args[4]
|
|
|
8 hotspotsfile = args[5]
|
|
14
|
9 aafile = args[6]
|
|
|
10 gene_identification_file= args[7]
|
|
|
11 output = args[8]
|
|
|
12 before.unique.file = args[9]
|
|
|
13 unmatchedfile = args[10]
|
|
|
14 method=args[11]
|
|
|
15 functionality=args[12]
|
|
|
16 unique.type=args[13]
|
|
|
17 filter.unique=args[14]
|
|
|
18 class.filter=args[15]
|
|
|
19 empty.region.filter=args[16]
|
|
0
|
20
|
|
|
21 summ = read.table(summaryfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="")
|
|
|
22 sequences = read.table(sequencesfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="")
|
|
|
23 mutationanalysis = read.table(mutationanalysisfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="")
|
|
|
24 mutationstats = read.table(mutationstatsfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="")
|
|
|
25 hotspots = read.table(hotspotsfile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="")
|
|
14
|
26 AAs = read.table(aafile, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="")
|
|
0
|
27 gene_identification = read.table(gene_identification_file, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="")
|
|
|
28
|
|
47
|
29 fix_column_names = function(df){
|
|
|
30 if("V.DOMAIN.Functionality" %in% names(df)){
|
|
|
31 names(df)[names(df) == "V.DOMAIN.Functionality"] = "Functionality"
|
|
|
32 print("found V.DOMAIN.Functionality, changed")
|
|
|
33 }
|
|
|
34 if("V.DOMAIN.Functionality.comment" %in% names(df)){
|
|
|
35 names(df)[names(df) == "V.DOMAIN.Functionality.comment"] = "Functionality.comment"
|
|
|
36 print("found V.DOMAIN.Functionality.comment, changed")
|
|
|
37 }
|
|
|
38 return(df)
|
|
|
39 }
|
|
|
40
|
|
|
41 summ = fix_column_names(summ)
|
|
|
42 sequences = fix_column_names(sequences)
|
|
|
43 mutationanalysis = fix_column_names(mutationanalysis)
|
|
|
44 mutationstats = fix_column_names(mutationstats)
|
|
|
45 hotspots = fix_column_names(hotspots)
|
|
|
46 AAs = fix_column_names(AAs)
|
|
|
47
|
|
0
|
48 if(method == "blastn"){
|
|
10
|
49 #"qseqid\tsseqid\tpident\tlength\tmismatch\tgapopen\tqstart\tqend\tsstart\tsend\tevalue\tbitscore"
|
|
0
|
50 gene_identification = gene_identification[!duplicated(gene_identification$qseqid),]
|
|
|
51 ref_length = data.frame(sseqid=c("ca1", "ca2", "cg1", "cg2", "cg3", "cg4", "cm"), ref.length=c(81,81,141,141,141,141,52))
|
|
|
52 gene_identification = merge(gene_identification, ref_length, by="sseqid", all.x=T)
|
|
|
53 gene_identification$chunk_hit_percentage = (gene_identification$length / gene_identification$ref.length) * 100
|
|
|
54 gene_identification = gene_identification[,c("qseqid", "chunk_hit_percentage", "pident", "qstart", "sseqid")]
|
|
|
55 colnames(gene_identification) = c("Sequence.ID", "chunk_hit_percentage", "nt_hit_percentage", "start_locations", "best_match")
|
|
|
56 }
|
|
|
57
|
|
47
|
58 #print("Summary analysis files columns")
|
|
|
59 #print(names(summ))
|
|
41
|
60
|
|
|
61
|
|
|
62
|
|
0
|
63 input.sequence.count = nrow(summ)
|
|
|
64 print(paste("Number of sequences in summary file:", input.sequence.count))
|
|
|
65
|
|
|
66 filtering.steps = data.frame(character(0), numeric(0))
|
|
|
67
|
|
|
68 filtering.steps = rbind(filtering.steps, c("Input", input.sequence.count))
|
|
|
69
|
|
|
70 filtering.steps[,1] = as.character(filtering.steps[,1])
|
|
|
71 filtering.steps[,2] = as.character(filtering.steps[,2])
|
|
|
72 #filtering.steps[,3] = as.numeric(filtering.steps[,3])
|
|
|
73
|
|
40
|
74 #print("summary files columns")
|
|
|
75 #print(names(summ))
|
|
38
|
76
|
|
0
|
77 summ = merge(summ, gene_identification, by="Sequence.ID")
|
|
|
78
|
|
|
79 summ = summ[summ$Functionality != "No results",]
|
|
|
80
|
|
|
81 print(paste("Number of sequences after 'No results' filter:", nrow(summ)))
|
|
|
82
|
|
|
83 filtering.steps = rbind(filtering.steps, c("After 'No results' filter", nrow(summ)))
|
|
|
84
|
|
|
85 if(functionality == "productive"){
|
|
|
86 summ = summ[summ$Functionality == "productive (see comment)" | summ$Functionality == "productive",]
|
|
|
87 } else if (functionality == "unproductive"){
|
|
|
88 summ = summ[summ$Functionality == "unproductive (see comment)" | summ$Functionality == "unproductive",]
|
|
|
89 } else if (functionality == "remove_unknown"){
|
|
|
90 summ = summ[summ$Functionality != "No results" & summ$Functionality != "unknown (see comment)" & summ$Functionality != "unknown",]
|
|
|
91 }
|
|
|
92
|
|
1
|
93 print(paste("Number of sequences after functionality filter:", nrow(summ)))
|
|
0
|
94
|
|
1
|
95 filtering.steps = rbind(filtering.steps, c("After functionality filter", nrow(summ)))
|
|
0
|
96
|
|
47
|
97 if(FALSE){ #to speed up debugging
|
|
|
98 set.seed(1)
|
|
|
99 summ = summ[sample(nrow(summ), floor(nrow(summ) * 0.05)),]
|
|
|
100 print(paste("Number of sequences after sampling 5%:", nrow(summ)))
|
|
|
101
|
|
|
102 filtering.steps = rbind(filtering.steps, c("Number of sequences after sampling 5%", nrow(summ)))
|
|
|
103 }
|
|
|
104
|
|
41
|
105 print("mutation analysis files columns")
|
|
|
106 print(names(mutationanalysis[,!(names(mutationanalysis) %in% names(summ)[-1])]))
|
|
38
|
107
|
|
0
|
108 result = merge(summ, mutationanalysis[,!(names(mutationanalysis) %in% names(summ)[-1])], by="Sequence.ID")
|
|
|
109
|
|
|
110 print(paste("Number of sequences after merging with mutation analysis file:", nrow(result)))
|
|
|
111
|
|
47
|
112 #print("mutation stats files columns")
|
|
|
113 #print(names(mutationstats[,!(names(mutationstats) %in% names(result)[-1])]))
|
|
38
|
114
|
|
0
|
115 result = merge(result, mutationstats[,!(names(mutationstats) %in% names(result)[-1])], by="Sequence.ID")
|
|
|
116
|
|
|
117 print(paste("Number of sequences after merging with mutation stats file:", nrow(result)))
|
|
|
118
|
|
41
|
119 print("hotspots files columns")
|
|
|
120 print(names(hotspots[,!(names(hotspots) %in% names(result)[-1])]))
|
|
38
|
121
|
|
0
|
122 result = merge(result, hotspots[,!(names(hotspots) %in% names(result)[-1])], by="Sequence.ID")
|
|
|
123
|
|
|
124 print(paste("Number of sequences after merging with hotspots file:", nrow(result)))
|
|
|
125
|
|
41
|
126 print("sequences files columns")
|
|
|
127 print(c("FR1.IMGT", "CDR1.IMGT", "FR2.IMGT", "CDR2.IMGT", "FR3.IMGT", "CDR3.IMGT"))
|
|
|
128
|
|
0
|
129 sequences = sequences[,c("Sequence.ID", "FR1.IMGT", "CDR1.IMGT", "FR2.IMGT", "CDR2.IMGT", "FR3.IMGT", "CDR3.IMGT")]
|
|
|
130 names(sequences) = c("Sequence.ID", "FR1.IMGT.seq", "CDR1.IMGT.seq", "FR2.IMGT.seq", "CDR2.IMGT.seq", "FR3.IMGT.seq", "CDR3.IMGT.seq")
|
|
|
131 result = merge(result, sequences, by="Sequence.ID", all.x=T)
|
|
|
132
|
|
41
|
133 print("sequences files columns")
|
|
|
134 print("CDR3.IMGT")
|
|
|
135
|
|
14
|
136 AAs = AAs[,c("Sequence.ID", "CDR3.IMGT")]
|
|
|
137 names(AAs) = c("Sequence.ID", "CDR3.IMGT.AA")
|
|
|
138 result = merge(result, AAs, by="Sequence.ID", all.x=T)
|
|
|
139
|
|
0
|
140 print(paste("Number of sequences in result after merging with sequences:", nrow(result)))
|
|
|
141
|
|
|
142 result$VGene = gsub("^Homsap ", "", result$V.GENE.and.allele)
|
|
|
143 result$VGene = gsub("[*].*", "", result$VGene)
|
|
|
144 result$DGene = gsub("^Homsap ", "", result$D.GENE.and.allele)
|
|
|
145 result$DGene = gsub("[*].*", "", result$DGene)
|
|
|
146 result$JGene = gsub("^Homsap ", "", result$J.GENE.and.allele)
|
|
|
147 result$JGene = gsub("[*].*", "", result$JGene)
|
|
|
148
|
|
12
|
149 splt = strsplit(class.filter, "_")[[1]]
|
|
|
150 chunk_hit_threshold = as.numeric(splt[1])
|
|
|
151 nt_hit_threshold = as.numeric(splt[2])
|
|
|
152
|
|
|
153 higher_than=(result$chunk_hit_percentage >= chunk_hit_threshold & result$nt_hit_percentage >= nt_hit_threshold)
|
|
|
154
|
|
|
155 if(!all(higher_than, na.rm=T)){ #check for no unmatched
|
|
|
156 result[!higher_than,"best_match"] = paste("unmatched,", result[!higher_than,"best_match"])
|
|
|
157 }
|
|
|
158
|
|
|
159 if(class.filter == "101_101"){
|
|
|
160 result$best_match = "all"
|
|
|
161 }
|
|
|
162
|
|
|
163 write.table(x=result, file=gsub("merged.txt$", "before_filters.txt", output), sep="\t",quote=F,row.names=F,col.names=T)
|
|
|
164
|
|
47
|
165 print(paste("Number of empty CDR1 sequences:", sum(result$CDR1.IMGT.seq == "", na.rm=T)))
|
|
|
166 print(paste("Number of empty FR2 sequences:", sum(result$FR2.IMGT.seq == "", na.rm=T)))
|
|
|
167 print(paste("Number of empty CDR2 sequences:", sum(result$CDR2.IMGT.seq == "", na.rm=T)))
|
|
|
168 print(paste("Number of empty FR3 sequences:", sum(result$FR3.IMGT.seq == "", na.rm=T)))
|
|
0
|
169
|
|
1
|
170 if(empty.region.filter == "leader"){
|
|
|
171 result = result[result$FR1.IMGT.seq != "" & result$CDR1.IMGT.seq != "" & result$FR2.IMGT.seq != "" & result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ]
|
|
|
172 } else if(empty.region.filter == "FR1"){
|
|
0
|
173 result = result[result$CDR1.IMGT.seq != "" & result$FR2.IMGT.seq != "" & result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ]
|
|
|
174 } else if(empty.region.filter == "CDR1"){
|
|
|
175 result = result[result$FR2.IMGT.seq != "" & result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ]
|
|
|
176 } else if(empty.region.filter == "FR2"){
|
|
|
177 result = result[result$CDR2.IMGT.seq != "" & result$FR3.IMGT.seq != "", ]
|
|
|
178 }
|
|
|
179
|
|
30
|
180 print(paste("After removal sequences that are missing a gene region:", nrow(result)))
|
|
|
181 filtering.steps = rbind(filtering.steps, c("After removal sequences that are missing a gene region", nrow(result)))
|
|
|
182
|
|
1
|
183 if(empty.region.filter == "leader"){
|
|
|
184 result = result[!(grepl("n|N", result$FR1.IMGT.seq) | grepl("n|N", result$FR2.IMGT.seq) | grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR1.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),]
|
|
|
185 } else if(empty.region.filter == "FR1"){
|
|
0
|
186 result = result[!(grepl("n|N", result$FR2.IMGT.seq) | grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR1.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),]
|
|
|
187 } else if(empty.region.filter == "CDR1"){
|
|
|
188 result = result[!(grepl("n|N", result$FR2.IMGT.seq) | grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),]
|
|
|
189 } else if(empty.region.filter == "FR2"){
|
|
2
|
190 result = result[!(grepl("n|N", result$FR3.IMGT.seq) | grepl("n|N", result$CDR2.IMGT.seq) | grepl("n|N", result$CDR3.IMGT.seq)),]
|
|
0
|
191 }
|
|
|
192
|
|
|
193 print(paste("Number of sequences in result after n filtering:", nrow(result)))
|
|
|
194 filtering.steps = rbind(filtering.steps, c("After N filter", nrow(result)))
|
|
|
195
|
|
|
196 cleanup_columns = c("FR1.IMGT.Nb.of.mutations",
|
|
|
197 "CDR1.IMGT.Nb.of.mutations",
|
|
|
198 "FR2.IMGT.Nb.of.mutations",
|
|
|
199 "CDR2.IMGT.Nb.of.mutations",
|
|
|
200 "FR3.IMGT.Nb.of.mutations")
|
|
|
201
|
|
|
202 for(col in cleanup_columns){
|
|
|
203 result[,col] = gsub("\\(.*\\)", "", result[,col])
|
|
|
204 result[,col] = as.numeric(result[,col])
|
|
|
205 result[is.na(result[,col]),] = 0
|
|
|
206 }
|
|
|
207
|
|
5
|
208 write.table(result, before.unique.file, sep="\t", quote=F,row.names=F,col.names=T)
|
|
|
209
|
|
|
210 if(filter.unique != "no"){
|
|
|
211 clmns = names(result)
|
|
|
212 if(empty.region.filter == "leader"){
|
|
|
213 result$unique.def = paste(result$FR1.IMGT.seq, result$CDR1.IMGT.seq, result$FR2.IMGT.seq, result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq)
|
|
|
214 } else if(empty.region.filter == "FR1"){
|
|
|
215 result$unique.def = paste(result$CDR1.IMGT.seq, result$FR2.IMGT.seq, result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq)
|
|
|
216 } else if(empty.region.filter == "CDR1"){
|
|
|
217 result$unique.def = paste(result$FR2.IMGT.seq, result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq)
|
|
|
218 } else if(empty.region.filter == "FR2"){
|
|
|
219 result$unique.def = paste(result$CDR2.IMGT.seq, result$FR3.IMGT.seq, result$CDR3.IMGT.seq)
|
|
|
220 }
|
|
12
|
221
|
|
15
|
222 if(filter.unique == "remove"){
|
|
5
|
223 result = result[duplicated(result$unique.def) | duplicated(result$unique.def, fromLast=T),]
|
|
|
224 }
|
|
40
|
225
|
|
15
|
226 result$unique.def = paste(result$unique.def, gsub(",.*", "", result$best_match)) #keep the unique sequences that are in multiple classes, gsub so the unmatched don't have a class after it
|
|
|
227
|
|
|
228 result = result[!duplicated(result$unique.def),]
|
|
1
|
229 }
|
|
|
230
|
|
5
|
231 write.table(result, gsub("before_unique_filter.txt", "after_unique_filter.txt", before.unique.file), sep="\t", quote=F,row.names=F,col.names=T)
|
|
|
232
|
|
|
233 filtering.steps = rbind(filtering.steps, c("After filter unique sequences", nrow(result)))
|
|
|
234
|
|
40
|
235 print(paste("Number of sequences in result after unique filtering:", nrow(result)))
|
|
|
236
|
|
1
|
237 if(nrow(summ) == 0){
|
|
|
238 stop("No data remaining after filter")
|
|
|
239 }
|
|
|
240
|
|
13
|
241 result$best_match_class = gsub(",.*", "", result$best_match) #gsub so the unmatched don't have a class after it
|
|
|
242
|
|
40
|
243 #result$past = ""
|
|
|
244 #cls = unlist(strsplit(unique.type, ","))
|
|
|
245 #for (i in 1:nrow(result)){
|
|
|
246 # result[i,"past"] = paste(result[i,cls], collapse=":")
|
|
|
247 #}
|
|
1
|
248
|
|
47
|
249
|
|
|
250
|
|
40
|
251 result$past = do.call(paste, c(result[unlist(strsplit(unique.type, ","))], sep = ":"))
|
|
15
|
252
|
|
|
253 result.matched = result[!grepl("unmatched", result$best_match),]
|
|
|
254 result.unmatched = result[grepl("unmatched", result$best_match),]
|
|
|
255
|
|
|
256 result = rbind(result.matched, result.unmatched)
|
|
|
257
|
|
1
|
258 result = result[!(duplicated(result$past)), ]
|
|
|
259
|
|
13
|
260 result = result[,!(names(result) %in% c("past", "best_match_class"))]
|
|
1
|
261
|
|
|
262 print(paste("Number of sequences in result after", unique.type, "filtering:", nrow(result)))
|
|
|
263
|
|
|
264 filtering.steps = rbind(filtering.steps, c("After remove duplicates based on filter", nrow(result)))
|
|
0
|
265
|
|
10
|
266 unmatched = result[grepl("^unmatched", result$best_match),c("Sequence.ID", "chunk_hit_percentage", "nt_hit_percentage", "start_locations", "best_match")]
|
|
|
267
|
|
0
|
268 print(paste("Number of rows in result:", nrow(result)))
|
|
|
269 print(paste("Number of rows in unmatched:", nrow(unmatched)))
|
|
|
270
|
|
|
271 matched.sequences = result[!grepl("^unmatched", result$best_match),]
|
|
|
272
|
|
|
273 write.table(x=matched.sequences, file=gsub("merged.txt$", "filtered.txt", output), sep="\t",quote=F,row.names=F,col.names=T)
|
|
|
274
|
|
|
275 matched.sequences.count = nrow(matched.sequences)
|
|
|
276 unmatched.sequences.count = sum(grepl("^unmatched", result$best_match))
|
|
|
277
|
|
|
278 filtering.steps = rbind(filtering.steps, c("Number of matched sequences", matched.sequences.count))
|
|
|
279 filtering.steps = rbind(filtering.steps, c("Number of unmatched sequences", unmatched.sequences.count))
|
|
|
280 filtering.steps[,2] = as.numeric(filtering.steps[,2])
|
|
|
281 filtering.steps$perc = round(filtering.steps[,2] / input.sequence.count * 100, 2)
|
|
|
282
|
|
|
283 write.table(x=filtering.steps, file=gsub("unmatched", "filtering_steps", unmatchedfile), sep="\t",quote=F,row.names=F,col.names=F)
|
|
|
284
|
|
|
285 write.table(x=result, file=output, sep="\t",quote=F,row.names=F,col.names=T)
|
|
|
286 write.table(x=unmatched, file=unmatchedfile, sep="\t",quote=F,row.names=F,col.names=T)
|
