| 1 |
// *************************************************************************** |
| 2 |
// FastaIndex.cpp (c) 2010 Erik Garrison <[email protected]> |
| 3 |
// Marth Lab, Department of Biology, Boston College |
| 4 |
// All rights reserved. |
| 5 |
// --------------------------------------------------------------------------- |
| 6 |
// Last modified: 9 February 2010 (EG) |
| 7 |
// --------------------------------------------------------------------------- |
| 8 |
|
| 9 |
#include "FBFasta.h" |
| 10 |
|
| 11 |
FB::FastaIndexEntry::FastaIndexEntry(string name, int length, long long offset, int line_blen, int line_len) |
| 12 |
: name(name) |
| 13 |
, length(length) |
| 14 |
, offset(offset) |
| 15 |
, line_blen(line_blen) |
| 16 |
, line_len(line_len) |
| 17 |
{} |
| 18 |
|
| 19 |
FB::FastaIndexEntry::FastaIndexEntry(void) // empty constructor |
| 20 |
{ clear(); } |
| 21 |
|
| 22 |
FB::FastaIndexEntry::~FastaIndexEntry(void) |
| 23 |
{} |
| 24 |
|
| 25 |
void FB::FastaIndexEntry::clear(void) |
| 26 |
{ |
| 27 |
name = ""; |
| 28 |
length = 0; |
| 29 |
offset = -1; // no real offset will ever be below 0, so this allows us to |
| 30 |
// check if we have already recorded a real offset |
| 31 |
line_blen = 0; |
| 32 |
line_len = 0; |
| 33 |
} |
| 34 |
|
| 35 |
ostream& FB::operator<<(ostream& output, const FastaIndexEntry& e) { |
| 36 |
// just write the first component of the name, for compliance with other tools |
| 37 |
output << split(e.name, ' ').at(0) << "\t" << e.length << "\t" << e.offset << "\t" << |
| 38 |
e.line_blen << "\t" << e.line_len; |
| 39 |
return output; // for multiple << operators. |
| 40 |
} |
| 41 |
|
| 42 |
FB::FastaIndex::FastaIndex(void) |
| 43 |
{} |
| 44 |
|
| 45 |
void FB::FastaIndex::readIndexFile(string fname) { |
| 46 |
string line; |
| 47 |
long long linenum = 0; |
| 48 |
indexFile.open(fname.c_str(), ifstream::in); |
| 49 |
if (indexFile.is_open()) { |
| 50 |
while (getline (indexFile, line)) { |
| 51 |
++linenum; |
| 52 |
// the fai format defined in samtools is tab-delimited, every line being: |
| 53 |
// fai->name[i], (int)x.len, (long long)x.offset, (int)x.line_blen, (int)x.line_len |
| 54 |
vector<string> fields = split(line, '\t'); |
| 55 |
if (fields.size() == 5) { // if we don't get enough fields then there is a problem with the file |
| 56 |
// note that fields[0] is the sequence name |
| 57 |
char* end; |
| 58 |
string name = split(fields[0], " \t").at(0); // key by first token of name |
| 59 |
sequenceNames.push_back(name); |
| 60 |
this->insert(make_pair(name, FastaIndexEntry(fields[0], atoi(fields[1].c_str()), |
| 61 |
strtoll(fields[2].c_str(), &end, 10), |
| 62 |
atoi(fields[3].c_str()), |
| 63 |
atoi(fields[4].c_str())))); |
| 64 |
} else { |
| 65 |
cerr << "Warning: malformed fasta index file " << fname << |
| 66 |
"does not have enough fields @ line " << linenum << endl; |
| 67 |
cerr << line << endl; |
| 68 |
exit(1); |
| 69 |
} |
| 70 |
} |
| 71 |
} else { |
| 72 |
cerr << "could not open index file " << fname << endl; |
| 73 |
exit(1); |
| 74 |
} |
| 75 |
} |
| 76 |
|
| 77 |
// for consistency this should be a class method |
| 78 |
bool fastaIndexEntryCompare ( FB::FastaIndexEntry a, FB::FastaIndexEntry b) { return (a.offset<b.offset); } |
| 79 |
|
| 80 |
ostream& FB::operator<<(ostream& output, FastaIndex& fastaIndex) { |
| 81 |
vector<FastaIndexEntry> sortedIndex; |
| 82 |
for(vector<string>::const_iterator it = fastaIndex.sequenceNames.begin(); it != fastaIndex.sequenceNames.end(); ++it) |
| 83 |
{ |
| 84 |
sortedIndex.push_back(fastaIndex[*it]); |
| 85 |
} |
| 86 |
sort(sortedIndex.begin(), sortedIndex.end(), fastaIndexEntryCompare); |
| 87 |
for( vector<FastaIndexEntry>::iterator fit = sortedIndex.begin(); fit != sortedIndex.end(); ++fit) { |
| 88 |
output << *fit << endl; |
| 89 |
} |
| 90 |
return output; |
| 91 |
} |
| 92 |
|
| 93 |
void FB::FastaIndex::indexReference(string refname) { |
| 94 |
// overview: |
| 95 |
// for line in the reference fasta file |
| 96 |
// track byte offset from the start of the file |
| 97 |
// if line is a fasta header, take the name and dump the last sequnece to the index |
| 98 |
// if line is a sequence, add it to the current sequence |
| 99 |
//cerr << "indexing fasta reference " << refname << endl; |
| 100 |
string line; |
| 101 |
FastaIndexEntry entry; // an entry buffer used in processing |
| 102 |
entry.clear(); |
| 103 |
int line_length = 0; |
| 104 |
long long offset = 0; // byte offset from start of file |
| 105 |
long long line_number = 0; // current line number |
| 106 |
bool mismatchedLineLengths = false; // flag to indicate if our line length changes mid-file |
| 107 |
// this will be used to raise an error |
| 108 |
// if we have a line length change at |
| 109 |
// any line other than the last line in |
| 110 |
// the sequence |
| 111 |
bool emptyLine = false; // flag to catch empty lines, which we allow for |
| 112 |
// index generation only on the last line of the sequence |
| 113 |
ifstream refFile; |
| 114 |
refFile.open(refname.c_str()); |
| 115 |
if (refFile.is_open()) { |
| 116 |
while (getline(refFile, line)) { |
| 117 |
++line_number; |
| 118 |
line_length = line.length(); |
| 119 |
if (line[0] == ';') { |
| 120 |
// fasta comment, skip |
| 121 |
} else if (line[0] == '+') { |
| 122 |
// fastq quality header |
| 123 |
getline(refFile, line); |
| 124 |
line_length = line.length(); |
| 125 |
offset += line_length + 1; |
| 126 |
// get and don't handle the quality line |
| 127 |
getline(refFile, line); |
| 128 |
line_length = line.length(); |
| 129 |
} else if (line[0] == '>' || line[0] == '@') { // fasta /fastq header |
| 130 |
// if we aren't on the first entry, push the last sequence into the index |
| 131 |
if (entry.name != "") { |
| 132 |
mismatchedLineLengths = false; // reset line length error tracker for every new sequence |
| 133 |
emptyLine = false; |
| 134 |
flushEntryToIndex(entry); |
| 135 |
entry.clear(); |
| 136 |
} |
| 137 |
entry.name = line.substr(1, line_length - 1); |
| 138 |
} else { // we assume we have found a sequence line |
| 139 |
if (entry.offset == -1) // NB initially the offset is -1 |
| 140 |
entry.offset = offset; |
| 141 |
entry.length += line_length; |
| 142 |
if (entry.line_len) { |
| 143 |
//entry.line_len = entry.line_len ? entry.line_len : line_length + 1; |
| 144 |
if (mismatchedLineLengths || emptyLine) { |
| 145 |
if (line_length == 0) { |
| 146 |
emptyLine = true; // flag empty lines, raise error only if this is embedded in the sequence |
| 147 |
} else { |
| 148 |
if (emptyLine) { |
| 149 |
cerr << "ERROR: embedded newline"; |
| 150 |
} else { |
| 151 |
cerr << "ERROR: mismatched line lengths"; |
| 152 |
} |
| 153 |
cerr << " at line " << line_number << " within sequence " << entry.name << |
| 154 |
endl << "File not suitable for fasta index generation." << endl; |
| 155 |
exit(1); |
| 156 |
} |
| 157 |
} |
| 158 |
// this flag is set here and checked on the next line |
| 159 |
// because we may have reached the end of the sequence, in |
| 160 |
// which case a mismatched line length is OK |
| 161 |
if (entry.line_len != line_length + 1) { |
| 162 |
mismatchedLineLengths = true; |
| 163 |
if (line_length == 0) { |
| 164 |
emptyLine = true; // flag empty lines, raise error only if this is embedded in the sequence |
| 165 |
} |
| 166 |
} |
| 167 |
} else { |
| 168 |
entry.line_len = line_length + 1; // first line |
| 169 |
} |
| 170 |
entry.line_blen = entry.line_len - 1; |
| 171 |
} |
| 172 |
offset += line_length + 1; |
| 173 |
} |
| 174 |
// we've hit the end of the fasta file! |
| 175 |
// flush the last entry |
| 176 |
flushEntryToIndex(entry); |
| 177 |
} else { |
| 178 |
cerr << "could not open reference file " << refname << " for indexing!" << endl; |
| 179 |
exit(1); |
| 180 |
} |
| 181 |
} |
| 182 |
|
| 183 |
void FB::FastaIndex::flushEntryToIndex(FastaIndexEntry& entry) { |
| 184 |
string name = split(entry.name, " \t").at(0); // key by first token of name |
| 185 |
sequenceNames.push_back(name); |
| 186 |
this->insert(make_pair(name, FastaIndexEntry(entry.name, entry.length, |
| 187 |
entry.offset, entry.line_blen, |
| 188 |
entry.line_len))); |
| 189 |
|
| 190 |
} |
| 191 |
|
| 192 |
void FB::FastaIndex::writeIndexFile(string fname) { |
| 193 |
//cerr << "writing fasta index file " << fname << endl; |
| 194 |
ofstream file; |
| 195 |
file.open(fname.c_str()); |
| 196 |
if (file.is_open()) { |
| 197 |
file << *this; |
| 198 |
} else { |
| 199 |
cerr << "could not open index file " << fname << " for writing!" << endl; |
| 200 |
exit(1); |
| 201 |
} |
| 202 |
} |
| 203 |
|
| 204 |
FB::FastaIndex::~FastaIndex(void) { |
| 205 |
indexFile.close(); |
| 206 |
} |
| 207 |
|
| 208 |
FB::FastaIndexEntry FB::FastaIndex::entry(string name) { |
| 209 |
FastaIndex::iterator e = this->find(name); |
| 210 |
if (e == this->end()) { |
| 211 |
cerr << "unable to find FASTA index entry for '" << name << "'" << endl; |
| 212 |
exit(1); |
| 213 |
} else { |
| 214 |
return e->second; |
| 215 |
} |
| 216 |
} |
| 217 |
|
| 218 |
string FB::FastaIndex::indexFileExtension() { return ".fai"; } |
| 219 |
|
| 220 |
/* |
| 221 |
FastaReference::FastaReference(string reffilename) { |
| 222 |
} |
| 223 |
*/ |
| 224 |
|
| 225 |
void FB::FastaReference::open(string reffilename) { |
| 226 |
filename = reffilename; |
| 227 |
if (!(file = fopen(filename.c_str(), "r"))) { |
| 228 |
cerr << "could not open " << filename << endl; |
| 229 |
exit(1); |
| 230 |
} |
| 231 |
index = new FastaIndex(); |
| 232 |
struct stat stFileInfo; |
| 233 |
string indexFileName = filename + index->indexFileExtension(); |
| 234 |
// if we can find an index file, use it |
| 235 |
if(stat(indexFileName.c_str(), &stFileInfo) == 0) { |
| 236 |
index->readIndexFile(indexFileName); |
| 237 |
} else { // otherwise, read the reference and generate the index file in the cwd |
| 238 |
cerr << "index file " << indexFileName << " not found, generating..." << endl; |
| 239 |
index->indexReference(filename); |
| 240 |
index->writeIndexFile(indexFileName); |
| 241 |
} |
| 242 |
} |
| 243 |
|
| 244 |
FB::FastaReference::~FastaReference(void) { |
| 245 |
fclose(file); |
| 246 |
delete index; |
| 247 |
} |
| 248 |
|
| 249 |
string removeIupacBases(string& str) { |
| 250 |
const string validBases = "ATGCN"; |
| 251 |
size_t found = str.find_first_not_of(validBases); |
| 252 |
while (found != string::npos) { |
| 253 |
str[found] = 'N'; |
| 254 |
found = str.find_first_not_of(validBases, found + 1); |
| 255 |
} |
| 256 |
return str; |
| 257 |
} |
| 258 |
|
| 259 |
string FB::FastaReference::getRawSequence(string seqname) { |
| 260 |
FastaIndexEntry entry = index->entry(seqname); |
| 261 |
int newlines_in_sequence = entry.length / entry.line_blen; |
| 262 |
int seqlen = newlines_in_sequence + entry.length; |
| 263 |
char* seq = (char*) calloc (seqlen + 1, sizeof(char)); |
| 264 |
fseek64(file, entry.offset, SEEK_SET); |
| 265 |
size_t x = fread(seq, sizeof(char), seqlen, file); |
| 266 |
seq[seqlen] = '\0'; |
| 267 |
char* pbegin = seq; |
| 268 |
char* pend = seq + (seqlen/sizeof(char)); |
| 269 |
pend = remove(pbegin, pend, '\n'); |
| 270 |
pend = remove(pbegin, pend, '\0'); |
| 271 |
string s = seq; |
| 272 |
free(seq); |
| 273 |
s.resize((pend - pbegin)/sizeof(char)); |
| 274 |
return s; |
| 275 |
} |
| 276 |
|
| 277 |
string FB::FastaReference::getSequence(string seqname) { |
| 278 |
string u = uppercase(getRawSequence(seqname)); |
| 279 |
return removeIupacBases(u); |
| 280 |
} |
| 281 |
|
| 282 |
// TODO cleanup; odd function. use a map |
| 283 |
string FB::FastaReference::sequenceNameStartingWith(string seqnameStart) { |
| 284 |
try { |
| 285 |
return (*index)[seqnameStart].name; |
| 286 |
} catch (const exception& e) { |
| 287 |
cerr << e.what() << ": unable to find index entry for " << seqnameStart << endl; |
| 288 |
exit(1); |
| 289 |
} |
| 290 |
} |
| 291 |
|
| 292 |
string FB::FastaReference::getRawSubSequence(string seqname, int start, int length) { |
| 293 |
FastaIndexEntry entry = index->entry(seqname); |
| 294 |
length = min(length, entry.length - start); |
| 295 |
if (start < 0 || length < 1) { |
| 296 |
return ""; |
| 297 |
} |
| 298 |
// we have to handle newlines |
| 299 |
// approach: count newlines before start |
| 300 |
// count newlines by end of read |
| 301 |
// subtracting newlines before start find count of embedded newlines |
| 302 |
int newlines_before = start > 0 ? (start - 1) / entry.line_blen : 0; |
| 303 |
int newlines_by_end = (start + length - 1) / entry.line_blen; |
| 304 |
int newlines_inside = newlines_by_end - newlines_before; |
| 305 |
int seqlen = length + newlines_inside; |
| 306 |
char* seq = (char*) calloc (seqlen + 1, sizeof(char)); |
| 307 |
fseek64(file, (off_t) (entry.offset + newlines_before + start), SEEK_SET); |
| 308 |
size_t x = fread(seq, sizeof(char), (off_t) seqlen, file); |
| 309 |
seq[seqlen] = '\0'; |
| 310 |
char* pbegin = seq; |
| 311 |
char* pend = seq + (seqlen/sizeof(char)); |
| 312 |
pend = remove(pbegin, pend, '\n'); |
| 313 |
pend = remove(pbegin, pend, '\0'); |
| 314 |
string s = seq; |
| 315 |
free(seq); |
| 316 |
s.resize((pend - pbegin)/sizeof(char)); |
| 317 |
return s; |
| 318 |
} |
| 319 |
|
| 320 |
string FB::FastaReference::getSubSequence(string seqname, int start, int length) { |
| 321 |
string u = uppercase(getRawSubSequence(seqname, start, length)); |
| 322 |
return removeIupacBases(u); |
| 323 |
} |
| 324 |
|
| 325 |
long unsigned int FB::FastaReference::sequenceLength(string seqname) { |
| 326 |
FastaIndexEntry entry = index->entry(seqname); |
| 327 |
return entry.length; |
| 328 |
} |
| 329 |
|