-
Notifications
You must be signed in to change notification settings - Fork 1
/
generate_fragments_for_training.R
237 lines (195 loc) · 9.57 KB
/
generate_fragments_for_training.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
#######################################################################################
### ###
### Copyright (C) 2017 Pawel Krawczyk (p.krawczyk@ibb.waw.pl) ###
### ###
### This program is free software: you can redistribute it and/or modify ###
### it under the terms of the GNU General Public License as published by ###
### the Free Software Foundation, either version 3 of the License, or ###
### (at your option) any later version. ###
### ###
### This program is distributed in the hope that it will be useful, ###
### but WITHOUT ANY WARRANTY; without even the implied warranty of ###
### MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ###
### GNU General Public License for more details. ###
### ###
### You should have received a copy of the GNU General Public License ###
### along with this program. If not, see <http://www.gnu.org/licenses/>. ###
### ###
#######################################################################################
library(Biostrings)
library(data.table)
library(plyr)
library(ggplot2)
library(XML)
library(rentrez)
# define splitting parameters
SAMPLE_LENGTH <- 10000
FRACTION_COVERED_PLASMIDS <- 0.75
FRACTION_COVERED_CHROMOSOMES <- 0.05
make_sample <- function(x, sample_length, fraction_covered) {
#function generating random sequence fragments of given length
splitted_set <- DNAStringSet()
x_length <- width(x)
print(paste("Sequence of length:", x_length))
if (x_length <= SAMPLE_LENGTH) {
splitted_set <- c(splitted_set, x)
} else {
no_samples <- ceiling(x_length/sample_length * fraction_covered)
print(paste("dividing into", no_samples, "samples"))
sequence_starts <- sample(x_length - SAMPLE_LENGTH - 1, no_samples)
for (z in sequence_starts) {
temp_seq <- subseq(x, z, z + SAMPLE_LENGTH - 1)
splitted_set <- c(splitted_set, temp_seq)
}
}
return(splitted_set)
}
generate_class2 <- function(x) {
# function generating classes used for training, based on origin (plasmid/chromosome) and taxonomy (phylum)
# to be used with apply
element1 <- as.character(x[9])
plasmid <- x[5]
element1 <- gsub("[[:space:]]", "", element1)
element1 <- gsub(",", "", element1)
element1 <- gsub("-", "", element1)
if (!is.na(element1)) {
if (element1 == "") {
element1 <- "other"
}
if (element1 == "undef") {
element1 <- "other"
}
} else {
element1 <- "other"
}
final_class = paste(plasmid, element1, sep = ".")
print(final_class)
return(final_class)
}
generate_class3 <- function(x) {
# classes tax rank 7 - class 4 - plasm
element1 <- as.character(x[10])
plasmid <- x[5]
element1 <- gsub("[[:space:]]", "", element1)
element1 <- gsub(",", "", element1)
element1 <- gsub("-", "", element1)
if (!is.na(element1)) {
if (element1 == "") {
element1 <- "other"
}
if (element1 == "undef") {
element1 <- "other"
}
} else {
element1 <- "other"
}
final_class = paste(plasmid, element1, sep = ".")
print(final_class)
return(final_class)
}
filter_class2 <- function(x) {
# functions used to filter low count classes
min_class_count <- 100 #minimum number of sequences to include in class
class_get <- as.character(x)
class_return = class_get
if (levels_count[levels_count$factor == class_get, ]$count < min_class_count) {
if (grepl("plasmid", class_get)) {
class_return = "plasmid.other"
}
if (grepl("chromosom", class_get)) {
class_return = "chromosome.other"
}
}
return(class_return)
}
# read previously imported sequences
load("imported_sequences.RData")
#load taxonomy annotation data
# should be manually edited to correct wrong annotations and exclude unwanted sequences
annotation_data<-fread("refseq_annotation.tsv",data.table=F)
accessions_df <- data.frame(accessions = accessions_dedup)
# get_annotation_of_all_sequences in the next step include Archaea
all_seqs_acc_annotated <- merge(annotation_data, accessions_df,
by.x = "accession", by.y = "accessions")
accessions_mapping <- data.frame()
#DNAStringSet for storing splitted sequences:
all_sequences_split <- DNAStringSet()
#vector storing accessions of splitted fragments
all_sequences_split_acc <- c()
#Perform actual random splitting:
for (y in seq(1, length(all_sequences_dedup))) {
temp_seq <- all_sequences_dedup[y]
temp_acc <- accessions_dedup[y]
if (nrow(all_seqs_acc_annotated[all_seqs_acc_annotated$accession == temp_acc,
]) > 0) {
#check if given sequence is plasmid
isplasmid <- all_seqs_acc_annotated[all_seqs_acc_annotated$accession == temp_acc,
]$plasmid
if (isplasmid == "chromosome") {
print(isplasmid)
splitted_set <- make_sample(temp_seq, SAMPLE_LENGTH, FRACTION_COVERED_CHROMOSOMES)
} else {
splitted_set <- make_sample(temp_seq, SAMPLE_LENGTH, FRACTION_COVERED_PLASMIDS)
}
}
print(paste("processed", y, "out of", length(all_sequences_dedup), "sequences"))
all_sequences_split <- c(all_sequences_split, splitted_set)
no_splitted_seqs <- length(splitted_set)
accessions_splitted <- rep(temp_acc, no_splitted_seqs)
all_sequences_split_acc <- c(all_sequences_split_acc, accessions_splitted)
}
#Calculate kmer counts (kmers from 3 to 7):
kmer3<-trinucleotideFrequency(all_sequences_split)
kmer3_frame<-as.data.frame(kmer3)
kmer3_frame$accessions<-all_sequences_split_acc
kmer4<-oligonucleotideFrequency(all_sequences_split,4)
kmer4_frame<-as.data.frame(kmer4)
kmer4_frame$accessions<-all_sequences_split_acc
kmer5<-oligonucleotideFrequency(all_sequences_split,5)
kmer5_frame<-as.data.frame(kmer5)
kmer5_frame$accessions<-all_sequences_split_acc
kmer6<-oligonucleotideFrequency(all_sequences_split,6)
kmer6_frame<-as.data.frame(kmer6)
kmer6_frame$accessions<-all_sequences_split_acc
kmer7<-oligonucleotideFrequency(all_sequences_split,7)
kmer7_frame<-as.data.frame(kmer7)
kmer7_frame$accessions<-all_sequences_split_acc
#annotate kmer count tables:
kmer3_annotated<-merge(annotation_tax_edited_no_Archaea,kmer3_frame,by.x="accession",by.y="accessions")
kmer4_annotated<-merge(annotation_tax_edited_no_Archaea,kmer4_frame,by.x="accession",by.y="accessions")
kmer5_annotated<-merge(annotation_tax_edited_no_Archaea,kmer5_frame,by.x="accession",by.y="accessions")
kmer6_annotated<-merge(annotation_tax_edited_no_Archaea,kmer6_frame,by.x="accession",by.y="accessions")
kmer7_annotated<-merge(annotation_tax_edited_no_Archaea,kmer7_frame,by.x="accession",by.y="accessions")
#generate class (in form plasmid.phylum) for each sequence
annotation_temp<-kmer3_annotated[,c(1:14)]
tax_classes<-apply(annotation,1,function(x) generate_class2(x))
#get counts for each class
levels_count<-as.data.frame(table(tax_classes))
colnames(levels_count) = c('factor','count')
#filter classes by count
tax_classes_filtered<-sapply(tax_classes,function(x) filter_class2(x))
#convert class names to numeric values (which will be provided as an input to TensorFlow)
tax_classes_numeric<-as.numeric(as.factor(tax_classes_filtered))-1
#create class labels data frame - will be required for prediction
class_labels<-levels(as.factor(tax_classes_filtered))
labels_df_num <- as.numeric(as.factor(class_labels)) - 1
labels_df <- data.frame(id = labels_df_num, label = class_labels)
# write class labels to file:
write.table(labels_df,file="class_labels_phyla_df.tsv",sep="\t",row.names=F)
#add generated classes to kmer counts tables and remove other annotation data:
kmer3_annotated_tax_classes<-cbind(tax_classes_numeric,kmer3_annotated[,-c(1:14)])
colnames(kmer3_annotated_tax_classes)[1]<-'plasmid'
kmer4_annotated_tax_classes<-cbind(tax_classes_numeric,kmer4_annotated[,-c(1:14)])
colnames(kmer4_annotated_tax_classes)[1]<-'plasmid'
kmer5_annotated_tax_classes<-cbind(tax_classes_numeric,kmer5_annotated[,-c(1:14)])
colnames(kmer5_annotated_tax_classes)[1]<-'plasmid'
kmer6_annotated_tax_classes<-cbind(tax_classes_numeric,kmer6_annotated[,-c(1:14)])
colnames(kmer6_annotated_tax_classes)[1]<-'plasmid'
kmer7_annotated_tax_classes<-cbind(tax_classes_numeric,kmer7_annotated[,-c(1:14)])
colnames(kmer7_annotated_tax_classes)[1]<-'plasmid'
#Write kmer count data to files:
write.table(kmer3_annotated_tax_classes,file="kmer3_split_raw_counts_tax_classes_numeric_annotated_filtered_tax.tsv",sep="\t",row.names=F,quote=F)
write.table(kmer4_annotated_tax_classes,file="kmer4_split_raw_counts_tax_classes_numeric_annotated_filtered_tax.tsv",sep="\t",row.names=F,quote=F)
write.table(kmer5_annotated_tax_classes,file="kmer5_split_raw_counts_tax_classes_numeric_annotated_filtered_tax.tsv",sep="\t",row.names=F,quote=F)
write.table(kmer6_annotated_tax_classes,file="kmer6_split_raw_counts_tax_classes_numeric_annotated_filtered_tax.tsv",sep="\t",row.names=F,quote=F)
write.table(kmer7_annotated_tax_classes,file="kmer7_split_raw_counts_tax_classes_numeric_annotated_filtered_tax.tsv",sep="\t",row.names=F,quote=F)