int main(int argc, char* argv[]) {
if (argc < 2) {
printf("Usage: PROBer-build-reference refName [--gtf gtfF] [--mapping mappingF] [--allele-specific] [--files num_of_files file_1 file_2 ...] [--n2g-index] [-q]\n");
exit(-1);
}
hasGTF = false;
mappingType = 0;
n2g_idx = false;
int argpos = 2;
while (argpos < argc) {
if (!strcmp(argv[argpos], "--gtf")) {
hasGTF = true;
strcpy(gtfF, argv[++argpos]);
}
if (!strcmp(argv[argpos], "--mapping")) {
mappingType = 1;
mappingPos = ++argpos;
}
if (!strcmp(argv[argpos], "--allele-specific")) mappingType = 2;
if (!strcmp(argv[argpos], "--files")) {
num_files = atoi(argv[++argpos]);
file_pos = argpos + 1; // the position in argv for the first file
argpos += num_files;
}
if (!strcmp(argv[argpos], "--n2g-index")) n2g_idx = true;
if (!strcmp(argv[argpos], "-q")) verbose = false;
++argpos;
}
if (mappingType > 0) loadMappingInfo(mappingType, argv[mappingPos]);
ifstream fin;
string line, gseq, tseq; // gseq, genomic sequence; tseq, transcript sequence
string seqname, gene_id, transcript_id;
if (hasGTF) {
transcripts.setType(0);
assert(mappingType < 2);
parse_gtf_file(gtfF);
M = transcripts.getM();
general_assert(M > 0, "The reference contains no transcripts!");
seqs.assign(M + 1, "");
chrvec.clear();
for (int i = 0; i < num_files; ++i, ++file_pos) {
fin.open(argv[file_pos]);
general_assert(fin.is_open(), "Cannot open " + cstrtos(argv[file_pos]) + "! It may not exist.");
getline(fin, line);
while ((fin) && (line[0] == '>')) {
istringstream strin(line.substr(1));
strin>>seqname;
gseq = "";
while((getline(fin, line)) && (line[0] != '>')) {
gseq += line;
}
assert(gseq.length() > 0);
sn2tr_iter = sn2tr.find(seqname);
if (sn2tr_iter == sn2tr.end()) continue;
chrvec.push_back(ChrInfo(seqname, gseq.length()));
vector<int>& vec = sn2tr_iter->second;
int s = vec.size();
for (int j = 0; j < s; ++j) {
assert(vec[j] > 0 && vec[j] <= M);
transcripts.getTranscriptAt(vec[j]).extractSeq(gseq, seqs[vec[j]]);
}
}
fin.close();
if (verbose) { printf("%s is processed!\n", argv[file_pos]); }
}
sort(chrvec.begin(), chrvec.end());
// Shrink and build up Refs
int curp = 0;
for (int i = 1; i <= M; ++i) {
const Transcript& transcript = transcripts.getTranscriptAt(i);
if (seqs[i] == "")
printf("Warning: Cannot extract transcript %s because the chromosome it locates -- %s -- is absent!\n", transcript.getTranscriptID().c_str(), transcript.getSeqName().c_str());
else {
refs.addRef(transcript.getTranscriptID(), seqs[i]); // insert RefSeqs
++curp;
transcripts.move(i, curp);
}
}
printf("%d transcripts are extracted and %d transcripts are omitted.\n", curp, M - curp);
transcripts.setM(curp);
M = transcripts.getM();
general_assert(M > 0, "The reference contains no transcripts!");
assert(refs.getM() == M);
}
else {
void parse_gtf_file(char* gtfF) {
ifstream fin(gtfF);
string line, curgid, tid, gid; // curgid: current gene id;
GTFItem item;
if (!fin.is_open()) { fprintf(stderr, "Cannot open %s! It may not exist.\n", gtfF); exit(-1); }
int cnt = 0;
items.clear();
while (getline(fin, line)) {
if (line[0] == '#') continue; // if this line is comment, jump it
item.parse(line);
string feature = item.getFeature();
if (feature == "exon") {
if (item.getStart() > item.getEnd()) {
fprintf(stderr, "Warning: exon's start position is larger than its end position! This exon is discarded.\n");
fprintf(stderr, "\t%s\n\n", line.c_str());
}
else if (item.getStart() < 1) {
fprintf(stderr, "Warning: exon's start position is less than 1! This exon is discarded.\n");
fprintf(stderr, "\t%s\n\n", line.c_str());
}
else {
if (hasMappingFile) {
tid = item.getTranscriptID();
mi_iter = mi_table.find(tid);
if (mi_iter == mi_table.end()) {
fprintf(stderr, "Mapping Info is not correct, cannot find %s's gene_id!\n", tid.c_str());
exit(-1);
}
//assert(iter != table.end());
gid = mi_iter->second;
item.setGeneID(gid);
}
items.push_back(item);
}
}
++cnt;
if (verbose && cnt % 200000 == 0) { printf("Parsed %d lines\n", cnt); }
}
fin.close();
sort(items.begin(), items.end());
starts.clear();
sn2tr.clear();
curgid = "";
int sp = 0, ep; // start pointer, end pointer
int nItems = items.size();
while (sp < nItems) {
tid = items[sp].getTranscriptID();
gid = items[sp].getGeneID();
ep = sp + 1;
while (ep < nItems && items[ep].getTranscriptID() == tid) ep++;
ep--;
buildTranscript(sp, ep);
int sid = transcripts.getM();
const Transcript& transcript = transcripts.getTranscriptAt(sid);
if (curgid != gid) {
starts.push_back(sid);
curgid = gid;
}
iter = sn2tr.find(transcript.getSeqName());
if (iter == sn2tr.end()) {
vector<int> vec(1, sid);
sn2tr[transcript.getSeqName()] = vec;
}
else {
iter->second.push_back(sid);
}
sp = ep + 1;
}
M = transcripts.getM();
starts.push_back(M + 1);
items.clear();
if (M < 1) {
fprintf(stderr, "Number of transcripts in the reference is less than 1!\n");
exit(-1);
}
if (verbose) { printf("Parsing gtf File is done!\n"); }
}
void parse_gtf_file(char* gtfF) {
ifstream fin(gtfF);
string line, tid, gid;
GTFItem item;
general_assert(fin.is_open(), "Cannot open " + cstrtos(gtfF) + "! It may not exist.");
int cnt = 0;
items.clear();
while (getline(fin, line)) {
if (skip(line)) continue;
item.parse(line);
string feature = item.getFeature();
if (feature == "exon") {
if (item.getStart() > item.getEnd()) {
printf("Warning: exon's start position is larger than its end position! This exon is discarded.\n");
printf("\t%s\n\n", line.c_str());
}
else if (item.getStart() < 1) {
printf("Warning: exon's start position is less than 1! This exon is discarded.\n");
printf("\t%s\n\n", line.c_str());
}
else {
item.parseAttributes(line);
if (mappingType > 0) {
tid = item.getTranscriptID();
mi_iter = mi_table.find(tid);
general_assert(mi_iter != mi_table.end(), "Mapping Info is not correct, cannot find " + tid + "'s gene_id!");
gid = mi_iter->second;
item.setGeneID(gid);
}
items.push_back(item);
}
}
++cnt;
if (verbose && cnt % 200000 == 0) { printf("Parsed %d lines\n", cnt); }
}
fin.close();
sort(items.begin(), items.end());
int sp = 0, ep; // start pointer, end pointer
int nItems = items.size();
sn2tr.clear();
while (sp < nItems) {
tid = items[sp].getTranscriptID();
ep = sp + 1;
while (ep < nItems && items[ep].getTranscriptID() == tid) ++ep;
--ep;
buildTranscript(sp, ep);
int sid = transcripts.getM();
const Transcript& transcript = transcripts.getTranscriptAt(sid);
sn2tr_iter = sn2tr.find(transcript.getSeqName());
if (sn2tr_iter == sn2tr.end()) {
vector<int> vec(1, sid);
sn2tr[transcript.getSeqName()] = vec;
}
else {
sn2tr_iter->second.push_back(sid);
}
sp = ep + 1;
}
items.clear();
if (verbose) { printf("Parsing GTF File is done!\n"); }
}
int main(int argc, char* argv[]) {
if (argc < 5 || ((hasMappingFile = atoi(argv[3])) && argc < 6)) {
printf("Usage: synthesisRef refName quiet hasMappingFile<0,no;1,yes;2,allele-specific> [mappingFile] reference_file_1 [reference_file_2 ...]\n");
exit(-1);
}
verbose = !atoi(argv[2]);
if (hasMappingFile) { loadMappingInfo(hasMappingFile, argv[4]); }
// allele-specific
if (hasMappingFile == 2) { transcripts.setType(2); }
int start = hasMappingFile ? 5 : 4;
ifstream fin;
string line, gseq;
string seqname, gene_id, transcript_id;
vector<Interval> vec;
M = 0;
name2seq.clear();
for (int i = start; i < argc; i++) {
fin.open(argv[i]);
general_assert(fin.is_open(), "Cannot open " + cstrtos(argv[i]) + "! It may not exist.");
unsigned long int line_no = 0; //Keep track of file line number
getline(fin, line);
line_no += 1;
while ((fin) && (line[0] == '>')) {
istringstream strin(line.substr(1));
strin>>seqname;
gseq = "";
while((getline(fin, line)) && (line[0] != '>')) {
line_no += 1;
gseq += line;
}
int len = gseq.length();
assert(len > 0);
for (int j = 0; j < len; j++) gseq[j] = check(gseq[j],line_no);
name2seq[seqname] = gseq;
transcript_id = seqname;
gene_id = seqname;
if (hasMappingFile) {
mi_iter = mi_table.find(seqname);
general_assert(mi_iter != mi_table.end(), "Mapping Info is not correct, cannot find " + seqname + "'s gene_id!");
gene_id = mi_iter->second;
if (hasMappingFile == 2) {
mi_iter2 = mi_table2.find(seqname);
general_assert(mi_iter2 != mi_table2.end(), "Mapping Info is not correct, cannot find allele " + seqname + "'s transcript_id!");
transcript_id = mi_iter2->second;
}
}
vec.clear();
vec.push_back(Interval(1, len));
transcripts.add(Transcript(transcript_id, gene_id, seqname, '+', vec, ""));
++M;
if (verbose && M % 1000000 == 0) { printf("%d sequences are processed!\n", M); }
}
fin.close();
}
if (M < 1) {
fprintf(stderr, "Number of transcripts in the reference is less than 1!\n");
exit(-1);
}
assert(M == transcripts.getM());
transcripts.sort();
writeResults(hasMappingFile, argv[1]);
return 0;
}
