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