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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304
|
/*
* Copyright 2011, Ben Langmead <langmea@cs.jhu.edu>
*
* This file is part of Bowtie 2.
*
* Bowtie 2 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.
*
* Bowtie 2 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 Bowtie 2. If not, see <http://www.gnu.org/licenses/>.
*/
#include "pat.h"
/**
* Parse a name from fb_ and store in r. Assume that the next
* character obtained via fb_.get() is the first character of
* the sequence and the string stops at the next char upto (could
* be tab, newline, etc.).
*/
int QseqPatternSource::parseName(
Read& r, // buffer for mate 1
Read* r2, // buffer for mate 2 (NULL if mate2 is read separately)
bool append, // true -> append characters, false -> skip them
bool clearFirst, // clear the name buffer first
bool warnEmpty, // emit a warning if nothing was added to the name
bool useDefault, // if nothing is read, put readCnt_ as a default value
int upto) // stop parsing when we first reach character 'upto'
{
if(clearFirst) {
if(r2 != NULL) r2->name.clear();
r.name.clear();
}
while(true) {
int c;
if((c = fb_.get()) < 0) {
// EOF reached in the middle of the name
return -1;
}
if(c == '\n' || c == '\r') {
// EOL reached in the middle of the name
return -1;
}
if(c == upto) {
// Finished with field
break;
}
if(append) {
if(r2 != NULL) r2->name.append(c);
r.name.append(c);
}
}
// Set up a default name if one hasn't been set
if(r.name.empty() && useDefault && append) {
char cbuf[20];
itoa10(readCnt_, cbuf);
r.name.append(cbuf);
if(r2 != NULL) r2->name.append(cbuf);
}
if(r.name.empty() && warnEmpty) {
cerr << "Warning: read had an empty name field" << endl;
}
return (int)r.name.length();
}
/**
* Parse a single sequence from fb_ and store in r. Assume
* that the next character obtained via fb_.get() is the first
* character of the sequence and the sequence stops at the next
* char upto (could be tab, newline, etc.).
*/
int QseqPatternSource::parseSeq(
Read& r,
int& charsRead,
int& trim5,
char upto)
{
int begin = 0;
int c = fb_.get();
assert(c != upto);
r.patFw.clear();
r.color = gColor;
if(gColor) {
// NOTE: clearly this is not relevant for Illumina output, but
// I'm keeping it here in case there's some reason to put SOLiD
// data in this format in the future.
// This may be a primer character. If so, keep it in the
// 'primer' field of the read buf and parse the rest of the
// read without it.
c = toupper(c);
if(asc2dnacat[c] > 0) {
// First char is a DNA char
int c2 = toupper(fb_.peek());
// Second char is a color char
if(asc2colcat[c2] > 0) {
r.primer = c;
r.trimc = c2;
trim5 += 2; // trim primer and first color
}
}
if(c < 0) { return -1; }
}
while(c != upto) {
if(c == '.') c = 'N';
if(gColor) {
if(c >= '0' && c <= '4') c = "ACGTN"[(int)c - '0'];
}
if(isalpha(c)) {
assert_in(toupper(c), "ACGTN");
if(begin++ >= trim5) {
assert_neq(0, asc2dnacat[c]);
r.patFw.append(asc2dna[c]);
}
charsRead++;
}
if((c = fb_.get()) < 0) {
return -1;
}
}
r.patFw.trimEnd(gTrim3);
return (int)r.patFw.length();
}
/**
* Parse a single quality string from fb_ and store in r.
* Assume that the next character obtained via fb_.get() is
* the first character of the quality string and the string stops
* at the next char upto (could be tab, newline, etc.).
*/
int QseqPatternSource::parseQuals(
Read& r,
int charsRead,
int dstLen,
int trim5,
char& c2,
char upto = '\t',
char upto2 = -1)
{
int qualsRead = 0;
int c = 0;
if (intQuals_) {
// Probably not relevant
char buf[4096];
while (qualsRead < charsRead) {
qualToks_.clear();
if(!tokenizeQualLine(fb_, buf, 4096, qualToks_)) break;
for (unsigned int j = 0; j < qualToks_.size(); ++j) {
char c = intToPhred33(atoi(qualToks_[j].c_str()), solQuals_);
assert_geq(c, 33);
if (qualsRead >= trim5) {
r.qual.append(c);
}
++qualsRead;
}
} // done reading integer quality lines
if (charsRead > qualsRead) tooFewQualities(r.name);
} else {
// Non-integer qualities
while((qualsRead < dstLen + trim5) && c >= 0) {
c = fb_.get();
c2 = c;
if (c == ' ') wrongQualityFormat(r.name);
if(c < 0) {
// EOF occurred in the middle of a read - abort
return -1;
}
if(!isspace(c) && c != upto && (upto2 == -1 || c != upto2)) {
if (qualsRead >= trim5) {
c = charToPhred33(c, solQuals_, phred64Quals_);
assert_geq(c, 33);
r.qual.append(c);
}
qualsRead++;
} else {
break;
}
}
}
if(r.qual.length() < (size_t)dstLen) {
tooFewQualities(r.name);
}
// TODO: How to detect too many qualities??
r.qual.resize(dstLen);
while(c != -1 && c != upto && (upto2 == -1 || c != upto2)) {
c = fb_.get();
c2 = c;
}
return qualsRead;
}
/**
* Read another pattern from a Qseq input file.
*/
bool QseqPatternSource::read(
Read& r,
TReadId& rdid,
TReadId& endid,
bool& success,
bool& done)
{
r.reset();
r.color = gColor;
success = true;
done = false;
readCnt_++;
rdid = endid = readCnt_-1;
peekOverNewline(fb_);
fb_.resetLastN();
// 1. Machine name
if(parseName(r, NULL, true, true, true, false, '\t') == -1) BAIL_UNPAIRED();
assert_neq('\t', fb_.peek());
r.name.append('_');
// 2. Run number
if(parseName(r, NULL, true, false, true, false, '\t') == -1) BAIL_UNPAIRED();
assert_neq('\t', fb_.peek());
r.name.append('_');
// 3. Lane number
if(parseName(r, NULL, true, false, true, false, '\t') == -1) BAIL_UNPAIRED();
assert_neq('\t', fb_.peek());
r.name.append('_');
// 4. Tile number
if(parseName(r, NULL, true, false, true, false, '\t') == -1) BAIL_UNPAIRED();
assert_neq('\t', fb_.peek());
r.name.append('_');
// 5. X coordinate of spot
if(parseName(r, NULL, true, false, true, false, '\t') == -1) BAIL_UNPAIRED();
assert_neq('\t', fb_.peek());
r.name.append('_');
// 6. Y coordinate of spot
if(parseName(r, NULL, true, false, true, false, '\t') == -1) BAIL_UNPAIRED();
assert_neq('\t', fb_.peek());
r.name.append('_');
// 7. Index
if(parseName(r, NULL, true, false, true, false, '\t') == -1) BAIL_UNPAIRED();
assert_neq('\t', fb_.peek());
r.name.append('/');
// 8. Mate number
if(parseName(r, NULL, true, false, true, false, '\t') == -1) BAIL_UNPAIRED();
// Empty sequence??
if(fb_.peek() == '\t') {
// Get tab that separates seq from qual
ASSERT_ONLY(int c =) fb_.get();
assert_eq('\t', c);
assert_eq('\t', fb_.peek());
// Get tab that separates qual from filter
ASSERT_ONLY(c =) fb_.get();
assert_eq('\t', c);
// Next char is first char of filter flag
assert_neq('\t', fb_.peek());
fb_.resetLastN();
cerr << "Warning: skipping empty QSEQ read with name '" << r.name << "'" << endl;
} else {
assert_neq('\t', fb_.peek());
int charsRead = 0;
int mytrim5 = gTrim5;
// 9. Sequence
int dstLen = parseSeq(r, charsRead, mytrim5, '\t');
assert_neq('\t', fb_.peek());
if(dstLen < 0) BAIL_UNPAIRED();
char ct = 0;
// 10. Qualities
if(parseQuals(r, charsRead, dstLen, mytrim5, ct, '\t', -1) < 0) BAIL_UNPAIRED();
r.trimmed3 = gTrim3;
r.trimmed5 = mytrim5;
if(ct != '\t') {
cerr << "Error: QSEQ with name " << r.name << " did not have tab after qualities" << endl;
throw 1;
}
assert_eq(ct, '\t');
}
// 11. Filter flag
int filt = fb_.get();
if(filt == -1) BAIL_UNPAIRED();
r.filter = filt;
if(filt != '0' && filt != '1') {
// Bad value for filt
}
if(fb_.peek() != -1 && fb_.peek() != '\n') {
// Bad value right after the filt field
}
fb_.get();
r.readOrigBuf.install(fb_.lastN(), fb_.lastNLen());
fb_.resetLastN();
if(r.qual.length() < r.patFw.length()) {
tooFewQualities(r.name);
} else if(r.qual.length() > r.patFw.length()) {
tooManyQualities(r.name);
}
#ifndef NDEBUG
assert_eq(r.patFw.length(), r.qual.length());
for(size_t i = 0; i < r.qual.length(); i++) {
assert_geq((int)r.qual[i], 33);
}
#endif
return true;
}
|