Mercurial > repos > davidvanzessen > report_clonality_igg
comparison RScript.r @ 22:2555b94dbdb2 draft
Uploaded
author | davidvanzessen |
---|---|
date | Thu, 15 Jan 2015 09:16:19 -0500 |
parents | 2424111b9198 |
children | 5f0597a3fd8b |
comparison
equal
deleted
inserted
replaced
21:2424111b9198 | 22:2555b94dbdb2 |
---|---|
166 PRODFJ = merge(PRODFJ, Total, by.x='Sample', by.y='Sample', all.x=TRUE) | 166 PRODFJ = merge(PRODFJ, Total, by.x='Sample', by.y='Sample', all.x=TRUE) |
167 PRODFJ = ddply(PRODFJ, c("Sample", "Top.J.Gene"), summarise, relFreq= (Length*100 / Total)) | 167 PRODFJ = ddply(PRODFJ, c("Sample", "Top.J.Gene"), summarise, relFreq= (Length*100 / Total)) |
168 | 168 |
169 # ---------------------- Setting up the gene names for the different T/B, human/mouse and locus ---------------------- | 169 # ---------------------- Setting up the gene names for the different T/B, human/mouse and locus ---------------------- |
170 | 170 |
171 V = c("v.name\tchr.orderV\n") | 171 genes = read.table("genes.txt", sep="\t", header=TRUE, fill=T, comment.char="") |
172 D = c("v.name\tchr.orderD\n") | 172 |
173 J = c("v.name\tchr.orderJ\n") | 173 Vchain = genes[grepl(species, genes$Species) & genes$locus == locus & genes$region == "V",c("IMGT.GENE.DB", "chr.order")] |
174 | 174 colnames(Vchain) = c("v.name", "chr.orderV") |
175 if(species == "human"){ | 175 Dchain = genes[grepl(species, genes$Species) & genes$locus == locus & genes$region == "D",c("IMGT.GENE.DB", "chr.order")] |
176 if(locus == "trb"){ | 176 colnames(Dchain) = c("v.name", "chr.orderD") |
177 V = c("v.name\tchr.orderV\nTRBV2\t1\nTRBV3-1\t2\nTRBV4-1\t3\nTRBV5-1\t4\nTRBV6-1\t5\nTRBV4-2\t6\nTRBV6-2\t7\nTRBV4-3\t8\nTRBV6-3\t9\nTRBV7-2\t10\nTRBV6-4\t11\nTRBV7-3\t12\nTRBV9\t13\nTRBV10-1\t14\nTRBV11-1\t15\nTRBV10-2\t16\nTRBV11-2\t17\nTRBV6-5\t18\nTRBV7-4\t19\nTRBV5-4\t20\nTRBV6-6\t21\nTRBV5-5\t22\nTRBV7-6\t23\nTRBV5-6\t24\nTRBV6-8\t25\nTRBV7-7\t26\nTRBV6-9\t27\nTRBV7-8\t28\nTRBV5-8\t29\nTRBV7-9\t30\nTRBV13\t31\nTRBV10-3\t32\nTRBV11-3\t33\nTRBV12-3\t34\nTRBV12-4\t35\nTRBV12-5\t36\nTRBV14\t37\nTRBV15\t38\nTRBV16\t39\nTRBV18\t40\nTRBV19\t41\nTRBV20-1\t42\nTRBV24-1\t43\nTRBV25-1\t44\nTRBV27\t45\nTRBV28\t46\nTRBV29-1\t47\nTRBV30\t48") | 177 Jchain = genes[grepl(species, genes$Species) & genes$locus == locus & genes$region == "J",c("IMGT.GENE.DB", "chr.order")] |
178 D = c("v.name\tchr.orderD\nTRBD1\t1\nTRBD2\t2\n") | 178 colnames(Jchain) = c("v.name", "chr.orderJ") |
179 J = c("v.name\tchr.orderJ\nTRBJ1-1\t1\nTRBJ1-2\t2\nTRBJ1-3\t3\nTRBJ1-4\t4\nTRBJ1-5\t5\nTRBJ1-6\t6\nTRBJ2-1\t7\nTRBJ2-2\t8\nTRBJ2-3\t9\nTRBJ2-4\t10\nTRBJ2-5\t11\nTRBJ2-6\t12\nTRBJ2-7\t13") | |
180 } else if (locus == "tra"){ | |
181 V = c("v.name\tchr.orderVTRAV1-1\t1\nTRAV1-2\t2\nTRAV2\t3\nTRAV3\t4\nTRAV4\t5\nTRAV5\t6\nTRAV6\t7\nTRAV7\t8\nTRAV8-1\t9\nTRAV9-1\t10\nTRAV10\t11\nTRAV12-1\t12\nTRAV8-2\t13\nTRAV8-3\t14\nTRAV13-1\t15\nTRAV12-2\t16\nTRAV8-4\t17\nTRAV13-2\t18\nTRAV14/DV4\t19\nTRAV9-2\t20\nTRAV12-3\t21\nTRAV8-6\t22\nTRAV16\t23\nTRAV17\t24\nTRAV18\t25\nTRAV19\t26\nTRAV20\t27\nTRAV21\t28\nTRAV22\t29\nTRAV23/DV6\t30\nTRAV24\t31\nTRAV25\t32\nTRAV26-1\t33\nTRAV27\t34\nTRAV29/DV5\t35\nTRAV30\t36\nTRAV26-2\t37\nTRAV34\t38\nTRAV35\t39\nTRAV36/DV7\t40\nTRAV38-1\t41\nTRAV38-2/DV8\t42\nTRAV39\t43\nTRAV40\t44\nTRAV41\t45\n") | |
182 D = c("v.name\tchr.orderD\n") | |
183 J = c("v.name\tchr.orderJ\nTRAJ57\t1\nTRAJ56\t2\nTRAJ54\t3\nTRAJ53\t4\nTRAJ52\t5\nTRAJ50\t6\nTRAJ49\t7\nTRAJ48\t8\nTRAJ47\t9\nTRAJ46\t10\nTRAJ45\t11\nTRAJ44\t12\nTRAJ43\t13\nTRAJ42\t14\nTRAJ41\t15\nTRAJ40\t16\nTRAJ39\t17\nTRAJ38\t18\nTRAJ37\t19\nTRAJ36\t20\nTRAJ34\t21\nTRAJ33\t22\nTRAJ32\t23\nTRAJ31\t24\nTRAJ30\t25\nTRAJ29\t26\nTRAJ28\t27\nTRAJ27\t28\nTRAJ26\t29\nTRAJ24\t30\nTRAJ23\t31\nTRAJ22\t32\nTRAJ21\t33\nTRAJ20\t34\nTRAJ18\t35\nTRAJ17\t36\nTRAJ16\t37\nTRAJ15\t38\nTRAJ14\t39\nTRAJ13\t40\nTRAJ12\t41\nTRAJ11\t42\nTRAJ10\t43\nTRAJ9\t44\nTRAJ8\t45\nTRAJ7\t46\nTRAJ6\t47\nTRAJ5\t48\nTRAJ4\t49\nTRAJ3\t50") | |
184 } else if (locus == "trg"){ | |
185 V = c("v.name\tchr.orderV\nTRGV9\t1\nTRGV8\t2\nTRGV5\t3\nTRGV4\t4\nTRGV3\t5\nTRGV2\t6") | |
186 D = c("v.name\tchr.orderD\n") | |
187 J = c("v.name\tchr.orderJ\nTRGJ2\t1\nTRGJP2\t2\nTRGJ1\t3\nTRGJP1\t4") | |
188 } else if (locus == "trd"){ | |
189 V = c("v.name\tchr.orderV\nTRDV1\t1\nTRDV2\t2\nTRDV3\t3") | |
190 D = c("v.name\tchr.orderD\nTRDD1\t1\nTRDD2\t2\nTRDD3\t3") | |
191 J = c("v.name\tchr.orderJ\nTRDJ1\t1\nTRDJ4\t2\nTRDJ2\t3\nTRDJ3\t4") | |
192 } else if(locus == "igh"){ | |
193 V = c("v.name\tchr.orderV\nIGHV3-74\t1\nIGHV3-73\t2\nIGHV3-72\t3\nIGHV2-70\t4\nIGHV1-69D\t5\nIGHV1-69-2\t6\nIGHV2-70D\t7\nIGHV1-69\t8\nIGHV3-66\t9\nIGHV3-64\t10\nIGHV4-61\t11\nIGHV4-59\t12\nIGHV1-58\t13\nIGHV3-53\t14\nIGHV5-51\t15\nIGHV3-49\t16\nIGHV3-48\t17\nIGHV1-46\t18\nIGHV1-45\t19\nIGHV3-43\t20\nIGHV4-39\t21\nIGHV3-43D\t22\nIGHV4-38-2\t23\nIGHV4-34\t24\nIGHV3-33\t25\nIGHV4-31\t26\nIGHV3-30-5\t27\nIGHV4-30-4\t28\nIGHV3-30-3\t29\nIGHV4-30-2\t30\nIGHV4-30-1\t31\nIGHV3-30\t32\nIGHV4-28\t33\nIGHV2-26\t34\nIGHV1-24\t35\nIGHV3-23D\t36\nIGHV3-23\t37\nIGHV3-21\t38\nIGHV3-20\t39\nIGHV1-18\t40\nIGHV3-15\t41\nIGHV3-13\t42\nIGHV3-11\t43\nIGHV5-10-1\t44\nIGHV3-9\t45\nIGHV1-8\t46\nIGHV3-64D\t47\nIGHV3-7\t48\nIGHV2-5\t49\nIGHV7-4-1\t50\nIGHV4-4\t51\nIGHV1-3\t52\nIGHV1-2\t53\nIGHV6-1\t54") | |
194 D = c("v.name\tchr.orderD\nIGHD1-7\t1\nIGHD2-8\t2\nIGHD3-9\t3\nIGHD3-10\t4\nIGHD5-12\t5\nIGHD6-13\t6\nIGHD2-15\t7\nIGHD3-16\t8\nIGHD4-17\t9\nIGHD5-18\t10\nIGHD6-19\t11\nIGHD1-20\t12\nIGHD2-21\t13\nIGHD3-22\t14\nIGHD5-24\t15\nIGHD6-25\t16\nIGHD1-26\t17\nIGHD7-27\t18") | |
195 J = c("v.name\tchr.orderJ\nIGHJ1\t1\nIGHJ2\t2\nIGHJ3\t3\nIGHJ4\t4\nIGHJ5\t5\nIGHJ6\t6") | |
196 } else if (locus == "igk"){ | |
197 V = c("v.name\tchr.orderV\nIGKV3D-7\t1\nIGKV1D-8\t2\nIGKV1D-43\t3\nIGKV3D-11\t4\nIGKV1D-12\t5\nIGKV1D-13\t6\nIGKV3D-15\t7\nIGKV1D-16\t8\nIGKV1D-17\t9\nIGKV3D-20\t10\nIGKV2D-26\t11\nIGKV2D-28\t12\nIGKV2D-29\t13\nIGKV2D-30\t14\nIGKV1D-33\t15\nIGKV1D-39\t16\nIGKV2D-40\t17\nIGKV2-40\t18\nIGKV1-39\t19\nIGKV1-33\t20\nIGKV2-30\t21\nIGKV2-29\t22\nIGKV2-28\t23\nIGKV1-27\t24\nIGKV2-24\t25\nIGKV3-20\t26\nIGKV1-17\t27\nIGKV1-16\t28\nIGKV3-15\t29\nIGKV1-13\t30\nIGKV1-12\t31\nIGKV3-11\t32\nIGKV1-9\t33\nIGKV1-8\t34\nIGKV1-6\t35\nIGKV1-5\t36\nIGKV5-2\t37\nIGKV4-1\t38") | |
198 D = c("v.name\tchr.orderD\n") | |
199 J = c("v.name\tchr.orderJ\nIGKJ1\t1\nIGKJ2\t2\nIGKJ3\t3\nIGKJ4\t4\nIGKJ5\t5") | |
200 } else if (locus == "igl"){ | |
201 V = c("v.name\tchr.orderV\nIGLV4-69\t1\nIGLV8-61\t2\nIGLV4-60\t3\nIGLV6-57\t4\nIGLV5-52\t5\nIGLV1-51\t6\nIGLV9-49\t7\nIGLV1-47\t8\nIGLV7-46\t9\nIGLV5-45\t10\nIGLV1-44\t11\nIGLV7-43\t12\nIGLV1-41\t13\nIGLV1-40\t14\nIGLV5-39\t15\nIGLV5-37\t16\nIGLV1-36\t17\nIGLV3-27\t18\nIGLV3-25\t19\nIGLV2-23\t20\nIGLV3-22\t21\nIGLV3-21\t22\nIGLV3-19\t23\nIGLV2-18\t24\nIGLV3-16\t25\nIGLV2-14\t26\nIGLV3-12\t27\nIGLV2-11\t28\nIGLV3-10\t29\nIGLV3-9\t30\nIGLV2-8\t31\nIGLV4-3\t32\nIGLV3-1\t33") | |
202 D = c("v.name\tchr.orderD\n") | |
203 J = c("v.name\tchr.orderJ\nIGLJ1\t1\nIGLJ2\t2\nIGLJ3\t3\nIGLJ6\t4\nIGLJ7\t5") | |
204 } | |
205 } else if (species == "mouse"){ | |
206 if(locus == "trb"){ | |
207 V = c("v.name\tchr.orderV\nTRBV1\t1\nTRBV2\t2\nTRBV3\t3\nTRBV4\t4\nTRBV5\t5\nTRBV12-1\t6\nTRBV13-1\t7\nTRBV12-2\t8\nTRBV13-2\t9\nTRBV13-3\t10\nTRBV14\t11\nTRBV15\t12\nTRBV16\t13\nTRBV17\t14\nTRBV19\t15\nTRBV20\t16\nTRBV23\t17\nTRBV24\t18\nTRBV26\t19\nTRBV29\t20\nTRBV30\t21\nTRBV31\t22") | |
208 D = c("v.name\tchr.orderD\nTRBD1\t1\nTRBD2\t2") | |
209 J = c("v.name\tchr.orderJ\nTRBJ1-1\t1\nTRBJ1-2\t2\nTRBJ1-3\t3\nTRBJ1-4\t4\nTRBJ1-5\t5\nTRBJ2-1\t6\nTRBJ2-2\t7\nTRBJ2-3\t8\nTRBJ2-4\t9\nTRBJ2-5\t10\nTRBJ2-6\t11\nTRBJ2-7\t12") | |
210 } else if (locus == "tra"){ | |
211 cat("mouse tra not yet implemented") | |
212 } else if (locus == "trg"){ | |
213 cat("mouse trg not yet implemented") | |
214 } else if (locus == "trd"){ | |
215 cat("mouse trd not yet implemented") | |
216 } else if(locus == "igh"){ | |
217 cat("mouse igh not yet implemented") | |
218 } else if (locus == "igk"){ | |
219 cat("mouse igk not yet implemented") | |
220 } else if (locus == "igl"){ | |
221 cat("mouse igl not yet implemented") | |
222 } | |
223 } | |
224 | 179 |
225 useD = TRUE | 180 useD = TRUE |
226 if(species == "human" && locus == "tra"){ | 181 if(nrow(Dchain) == 0){ |
227 useD = FALSE | 182 useD = FALSE |
228 cat("No D Genes in this species/locus") | 183 cat("No D Genes in this species/locus") |
229 } | 184 } |
230 | 185 |
231 # ---------------------- load the gene names into a data.frame and merge with the frequency count ---------------------- | 186 # ---------------------- merge with the frequency count ---------------------- |
232 | 187 |
233 tcV = textConnection(V) | |
234 Vchain = read.table(tcV, sep="\t", header=TRUE) | |
235 PRODFV = merge(PRODFV, Vchain, by.x='Top.V.Gene', by.y='v.name', all.x=TRUE) | 188 PRODFV = merge(PRODFV, Vchain, by.x='Top.V.Gene', by.y='v.name', all.x=TRUE) |
236 close(tcV) | 189 |
237 | |
238 tcD = textConnection(D) | |
239 Dchain = read.table(tcD, sep="\t", header=TRUE) | |
240 PRODFD = merge(PRODFD, Dchain, by.x='Top.D.Gene', by.y='v.name', all.x=TRUE) | 190 PRODFD = merge(PRODFD, Dchain, by.x='Top.D.Gene', by.y='v.name', all.x=TRUE) |
241 close(tcD) | 191 |
242 | |
243 tcJ = textConnection(J) | |
244 Jchain = read.table(tcJ, sep="\t", header=TRUE) | |
245 PRODFJ = merge(PRODFJ, Jchain, by.x='Top.J.Gene', by.y='v.name', all.x=TRUE) | 192 PRODFJ = merge(PRODFJ, Jchain, by.x='Top.J.Gene', by.y='v.name', all.x=TRUE) |
246 close(tcJ) | |
247 | 193 |
248 # ---------------------- Create the V, D and J frequency plots and write the data.frame for every plot to a file ---------------------- | 194 # ---------------------- Create the V, D and J frequency plots and write the data.frame for every plot to a file ---------------------- |
249 | 195 |
250 pV = ggplot(PRODFV) | 196 pV = ggplot(PRODFV) |
251 pV = pV + geom_bar( aes( x=factor(reorder(Top.V.Gene, chr.orderV)), y=relFreq, fill=Sample), stat='identity', position="dodge") + theme(axis.text.x = element_text(angle = 90, hjust = 1)) | 197 pV = pV + geom_bar( aes( x=factor(reorder(Top.V.Gene, chr.orderV)), y=relFreq, fill=Sample), stat='identity', position="dodge") + theme(axis.text.x = element_text(angle = 90, hjust = 1)) |