void writeResults(char* refName) {
int s;
ofstream fout;
sprintf(groupF, "%s.grp", refName);
sprintf(tiF, "%s.ti", refName);
sprintf(refFastaF, "%s.transcripts.fa", refName);
sprintf(chromListF, "%s.chrlist", refName);
fout.open(groupF);
s = starts.size();
for (int i = 0; i < s; i++) fout<<starts[i]<<endl;
fout.close();
if (verbose) { printf("Group File is generated!\n"); }
transcripts.writeTo(tiF);
if (verbose) { printf("Transcript Information File is generated!\n"); }
fout.open(chromListF);
s = chrvec.size();
for (int i = 0; i < s; i++) {
fout<<chrvec[i].name<<'\t'<<chrvec[i].len<<endl;
}
fout.close();
if (verbose) { printf("Chromosome List File is generated!\n"); }
fout.open(refFastaF);
for (int i = 1; i <= M; i++) {
fout<<">"<<transcripts.getTranscriptAt(i).getTranscriptID()<<endl;
fout<<seqs[i]<<endl;
}
fout.close();
if (verbose) { printf("Extracted Sequences File is generated!\n"); }
}
void buildTranscript(int sp, int ep) {
int cur_s, cur_e; // current_start, current_end
string transcript_id = items[sp].getTranscriptID();
string gene_id = items[sp].getGeneID();
string gene_name = "", transcript_name = "";
char strand = items[sp].getStrand();
string seqname = items[sp].getSeqName();
string left = items[sp].getLeft();
vec.clear();
cur_s = cur_e = -1;
for (int i = sp; i <= ep; ++i) {
int start = items[i].getStart();
int end = items[i].getEnd();
general_assert(strand == items[i].getStrand(), "According to the GTF file given, a transcript has exons from different orientations!");
general_assert(seqname == items[i].getSeqName(), "According to the GTF file given, a transcript has exons on multiple chromosomes!");
if (items[i].getGeneName() != "") {
if (gene_name == "") gene_name = items[i].getGeneName();
else general_assert(gene_name == items[i].getGeneName(), "A transcript is associated with multiple gene names!");
}
if (items[i].getTranscriptName() != "") {
if (transcript_name == "") transcript_name = items[i].getTranscriptName();
else general_assert(transcript_name == items[i].getTranscriptName(), "A transcript is associated with multiple transcript names!");
}
if (cur_e + 1 < start) {
if (cur_s > 0) vec.push_back(Interval(cur_s, cur_e));
cur_s = start;
}
cur_e = (cur_e < end ? end : cur_e);
}
if (cur_s > 0) vec.push_back(Interval(cur_s, cur_e));
// if (gene_name != "") gene_id += "_" + gene_name;
// if (transcript_name != "") transcript_id += "_" + transcript_name;
transcripts.add(Transcript(transcript_id, gene_id, seqname, strand, vec, left));
}
bool buildTranscript(int sp, int ep) {
int cur_s, cur_e; // current_start, current_end
vector<Interval> vec;
string transcript_id = items[sp].getTranscriptID();
string gene_id = items[sp].getGeneID();
char strand = items[sp].getStrand();
string seqname = items[sp].getSeqName();
string left = items[sp].getLeft();
vec.clear();
cur_s = cur_e = -1;
for (int i = sp; i <= ep; i++) {
int start = items[i].getStart();
int end = items[i].getEnd();
if (strand != items[i].getStrand()) {
fprintf(stderr, "According to the GTF file given, a transcript has exons from different orientations!\n");
exit(-1);
}
if (seqname != items[i].getSeqName()) {
fprintf(stderr, "According to the GTF file given, a transcript has exons on multiple chromosomes!\n");
exit(-1);
}
if (cur_e + 1 < start) {
if (cur_s > 0) vec.push_back(Interval(cur_s, cur_e));
cur_s = start;
}
cur_e = (cur_e < end ? end : cur_e);
}
if (cur_s > 0) vec.push_back(Interval(cur_s, cur_e));
transcripts.add(Transcript(transcript_id, gene_id, seqname, strand, vec, left));
return true;
}
int main(int argc, char* argv[]) {
ifstream fin;
bool quiet = false;
if (argc < 5) {
printf("Usage : rsem-run-em refName read_type sampleName sampleToken [-p #Threads] [-b samInpType samInpF has_fn_list_? [fn_list]] [-q] [--gibbs-out] [--sampling]\n\n");
printf(" refName: reference name\n");
printf(" read_type: 0 single read without quality score; 1 single read with quality score; 2 paired-end read without quality score; 3 paired-end read with quality score.\n");
printf(" sampleName: sample's name, including the path\n");
printf(" sampleToken: sampleName excludes the path\n");
printf(" -p: number of threads which user wants to use. (default: 1)\n");
printf(" -b: produce bam format output file. (default: off)\n");
printf(" -q: set it quiet\n");
printf(" --gibbs-out: generate output file used by Gibbs sampler. (default: off)\n");
printf(" --sampling: sample each read from its posterior distribution when bam file is generated. (default: off)\n");
printf("// model parameters should be in imdName.mparams.\n");
exit(-1);
}
time_t a = time(NULL);
strcpy(refName, argv[1]);
read_type = atoi(argv[2]);
strcpy(outName, argv[3]);
sprintf(imdName, "%s.temp/%s", argv[3], argv[4]);
sprintf(statName, "%s.stat/%s", argv[3], argv[4]);
nThreads = 1;
genBamF = false;
bamSampling = false;
genGibbsOut = false;
pt_fn_list = pt_chr_list = NULL;
for (int i = 5; i < argc; i++) {
if (!strcmp(argv[i], "-p")) { nThreads = atoi(argv[i + 1]); }
if (!strcmp(argv[i], "-b")) {
genBamF = true;
inpSamType = argv[i + 1][0];
strcpy(inpSamF, argv[i + 2]);
if (atoi(argv[i + 3]) == 1) {
strcpy(fn_list, argv[i + 4]);
pt_fn_list = (char*)(&fn_list);
}
}
if (!strcmp(argv[i], "-q")) { quiet = true; }
if (!strcmp(argv[i], "--gibbs-out")) { genGibbsOut = true; }
if (!strcmp(argv[i], "--sampling")) { bamSampling = true; }
}
general_assert(nThreads > 0, "Number of threads should be bigger than 0!");
verbose = !quiet;
//basic info loading
sprintf(refF, "%s.seq", refName);
refs.loadRefs(refF);
M = refs.getM();
sprintf(groupF, "%s.grp", refName);
gi.load(groupF);
m = gi.getm();
sprintf(tiF, "%s.ti", refName);
transcripts.readFrom(tiF);
sprintf(cntF, "%s.cnt", statName);
fin.open(cntF);
general_assert(fin.is_open(), "Cannot open " + cstrtos(cntF) + "! It may not exist.");
fin>>N0>>N1>>N2>>N_tot;
fin.close();
general_assert(N1 > 0, "There are no alignable reads!");
if ((READ_INT_TYPE)nThreads > N1) nThreads = N1;
//set model parameters
mparams.M = M;
mparams.N[0] = N0; mparams.N[1] = N1; mparams.N[2] = N2;
mparams.refs = &refs;
sprintf(mparamsF, "%s.mparams", imdName);
fin.open(mparamsF);
general_assert(fin.is_open(), "Cannot open " + cstrtos(mparamsF) + "It may not exist.");
fin>> mparams.minL>> mparams.maxL>> mparams.probF;
int val; // 0 or 1 , for estRSPD
fin>>val;
mparams.estRSPD = (val != 0);
fin>> mparams.B>> mparams.mate_minL>> mparams.mate_maxL>> mparams.mean>> mparams.sd;
fin>> mparams.seedLen;
fin.close();
//run EM
switch(read_type) {
case 0 : EM<SingleRead, SingleHit, SingleModel>(); break;
case 1 : EM<SingleReadQ, SingleHit, SingleQModel>(); break;
case 2 : EM<PairedEndRead, PairedEndHit, PairedEndModel>(); break;
case 3 : EM<PairedEndReadQ, PairedEndHit, PairedEndQModel>(); break;
default : fprintf(stderr, "Unknown Read Type!\n"); exit(-1);
}
time_t b = time(NULL);
printTimeUsed(a, b, "EM.cpp");
return 0;
}
void writeResults(ModelType& model, double* counts) {
double denom;
char outF[STRLEN];
FILE *fo;
sprintf(modelF, "%s.model", statName);
model.write(modelF);
//calculate tau values
double *tau = new double[M + 1];
memset(tau, 0, sizeof(double) * (M + 1));
denom = 0.0;
for (int i = 1; i <= M; i++)
if (eel[i] >= EPSILON) {
tau[i] = theta[i] / eel[i];
denom += tau[i];
}
general_assert(denom > 0, "No alignable reads?!");
for (int i = 1; i <= M; i++) {
tau[i] /= denom;
}
//isoform level results
sprintf(outF, "%s.iso_res", imdName);
fo = fopen(outF, "w");
for (int i = 1; i <= M; i++) {
const Transcript& transcript = transcripts.getTranscriptAt(i);
fprintf(fo, "%s%c", transcript.getTranscriptID().c_str(), (i < M ? '\t' : '\n'));
}
for (int i = 1; i <= M; i++)
fprintf(fo, "%.2f%c", counts[i], (i < M ? '\t' : '\n'));
for (int i = 1; i <= M; i++)
fprintf(fo, "%.15g%c", tau[i], (i < M ? '\t' : '\n'));
for (int i = 1; i <= M; i++) {
const Transcript& transcript = transcripts.getTranscriptAt(i);
fprintf(fo, "%s%c", transcript.getGeneID().c_str(), (i < M ? '\t' : '\n'));
}
fclose(fo);
//gene level results
sprintf(outF, "%s.gene_res", imdName);
fo = fopen(outF, "w");
for (int i = 0; i < m; i++) {
const string& gene_id = transcripts.getTranscriptAt(gi.spAt(i)).getGeneID();
fprintf(fo, "%s%c", gene_id.c_str(), (i < m - 1 ? '\t' : '\n'));
}
for (int i = 0; i < m; i++) {
double sumC = 0.0; // sum of counts
int b = gi.spAt(i), e = gi.spAt(i + 1);
for (int j = b; j < e; j++) sumC += counts[j];
fprintf(fo, "%.2f%c", sumC, (i < m - 1 ? '\t' : '\n'));
}
for (int i = 0; i < m; i++) {
double sumT = 0.0; // sum of tau values
int b = gi.spAt(i), e = gi.spAt(i + 1);
for (int j = b; j < e; j++) sumT += tau[j];
fprintf(fo, "%.15g%c", sumT, (i < m - 1 ? '\t' : '\n'));
}
for (int i = 0; i < m; i++) {
int b = gi.spAt(i), e = gi.spAt(i + 1);
for (int j = b; j < e; j++) {
fprintf(fo, "%s%c", transcripts.getTranscriptAt(j).getTranscriptID().c_str(), (j < e - 1 ? ',' : (i < m - 1 ? '\t' :'\n')));
}
}
fclose(fo);
delete[] tau;
if (verbose) { printf("Expression Results are written!\n"); }
}
int main(int argc, char* argv[]) {
bool quiet = false;
if (argc < 6) {
printf("Usage : rsem-parse-alignments refName imdName statName alignFType('s' for sam, 'b' for bam) alignF [-t Type] [-l fn_list] [-tag tagName] [-q]\n");
exit(-1);
}
strcpy(fn_list, "");
read_type = 0;
if (argc > 6) {
for (int i = 6; i < argc; i++) {
if (!strcmp(argv[i], "-t")) {
read_type = atoi(argv[i + 1]);
}
if (!strcmp(argv[i], "-l")) {
strcpy(fn_list, argv[i + 1]);
}
if (!strcmp(argv[i], "-tag")) {
SamParser::setReadTypeTag(argv[i + 1]);
}
if (!strcmp(argv[i], "-q")) { quiet = true; }
}
}
verbose = !quiet;
sprintf(groupF, "%s.grp", argv[1]);
gi.load(groupF);
sprintf(tiF, "%s.ti", argv[1]);
transcripts.readFrom(tiF);
sprintf(datF, "%s.dat", argv[2]);
sprintf(cntF, "%s.cnt", argv[3]);
init(argv[2], argv[4][0], argv[5]);
hit_out.open(datF);
string firstLine(99, ' ');
firstLine.append(1, '\n'); //May be dangerous!
hit_out<<firstLine;
switch(read_type) {
case 0 : parseIt<SingleRead, SingleHit>(parser); break;
case 1 : parseIt<SingleReadQ, SingleHit>(parser); break;
case 2 : parseIt<PairedEndRead, PairedEndHit>(parser); break;
case 3 : parseIt<PairedEndReadQ, PairedEndHit>(parser); break;
}
hit_out.seekp(0, ios_base::beg);
hit_out<<N[1]<<" "<<nHits<<" "<<read_type;
hit_out.close();
//cntF for statistics of alignments file
ofstream fout(cntF);
fout<<N[0]<<" "<<N[1]<<" "<<N[2]<<" "<<(N[0] + N[1] + N[2])<<endl;
fout<<nUnique<<" "<<nMulti<<" "<<nIsoMulti<<endl;
fout<<nHits<<" "<<read_type<<endl;
fout<<"0\t"<<N[0]<<endl;
for (iter = counter.begin(); iter != counter.end(); iter++) {
fout<<iter->first<<'\t'<<iter->second<<endl;
}
fout<<"Inf\t"<<N[2]<<endl;
fout.close();
release();
if (verbose) { printf("Done!\n"); }
return 0;
}
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 writeToDisk(char* refName) {
ofstream fout;
sprintf(tiF, "%s.ti", refName);
transcripts.writeTo(tiF);
if (verbose) { printf("Transcript Information File is generated!\n"); }
sprintf(refFastaF, "%s.transcripts.fa", refName);
refs.writeTo(refFastaF);
sprintf(transListF, "%s.translist", refName);
refs.writeTransListTo(transListF);
sprintf(chromListF, "%s.chrlist", refName);
fout.open(chromListF);
for (int i = 0; i < (int)chrvec.size(); ++i)
fout<< chrvec[i].name<< '\t'<< chrvec[i].len<< endl;
fout.close();
if (verbose) { printf("Chromosome List File is generated!\n"); }
string cur_gene_id, cur_transcript_id, name;
vector<int> gi, gt, ta;
cur_gene_id = ""; gi.clear();
if (mappingType == 2) { cur_transcript_id = ""; gt.clear(); ta.clear(); }
for (int i = 1; i <= M; ++i) {
const Transcript& transcript = transcripts.getTranscriptAt(i);
if (cur_gene_id != transcript.getGeneID()) {
gi.push_back(i);
if (mappingType == 2) gt.push_back((int)ta.size());
cur_gene_id = transcript.getGeneID();
}
if ((mappingType == 2) && (cur_transcript_id != transcript.getTranscriptID())) {
ta.push_back(i);
cur_transcript_id = transcript.getTranscriptID();
}
}
gi.push_back(M + 1);
if (mappingType == 2) { gt.push_back((int)ta.size()); ta.push_back(M + 1); }
sprintf(groupF, "%s.grp", refName);
fout.open(groupF);
for (int i = 0; i < (int)gi.size(); ++i) fout<< gi[i]<< endl;
fout.close();
if (verbose) { printf("Group File is generated!\n"); }
if (mappingType == 2) {
sprintf(gtF, "%s.gt", refName);
fout.open(gtF);
for (int i = 0; i < (int)gt.size(); ++i) fout<< gt[i]<< endl;
fout.close();
sprintf(taF, "%s.ta", refName);
fout.open(taF);
for (int i = 0; i < (int)ta.size(); ++i) fout<< ta[i]<< endl;
fout.close();
if (verbose) { printf("Allele-specific group files are generated!\n"); }
}
if (n2g_idx) {
sprintf(n2g_idxF, "%s.n2g.idx.fa", refName);
fout.open(n2g_idxF);
for (int i = 1; i <= M; ++i)
fout<< '>'<< refs.getRef(i)->getName()<< endl<< n2g(refs.getRef(i)->getSeq())<< endl;
fout.close();
if (verbose) printf("%s is generated!\n", n2g_idxF);
}
}
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 < 6 || ((hasMappingFile = atoi(argv[4])) && argc < 7)) {
printf("Usage: rsem-extract-reference-transcripts refName quiet gtfF hasMappingFile [mappingFile] chromosome_file_1 [chromosome_file_2 ...]\n");
exit(-1);
}
verbose = !atoi(argv[2]);
if (hasMappingFile) {
loadMappingInfo(argv[5]);
}
parse_gtf_file(argv[3]);
ifstream fin;
string line, gseq, seqname;
chrvec.clear();
seqs.clear();
seqs.resize(M + 1, "");
int start = hasMappingFile ? 6 : 5;
for (int i = start; i < argc; i++) {
fin.open(argv[i]);
if (!fin.is_open()) { fprintf(stderr, "Cannot open %s! It may not exist.\n", argv[i]); exit(-1); }
getline(fin, line);
while ((fin) && (line[0] == '>')) {
istringstream strin(line.substr(1));
strin>>seqname;
gseq = "";
while((getline(fin, line)) && (line[0] != '>')) {
gseq += line;
}
size_t len = gseq.length();
assert(len > 0);
for (size_t j = 0; j < len; j++) gseq[j] = check(gseq[j]);
iter = sn2tr.find(seqname);
if (iter == sn2tr.end()) continue;
chrvec.push_back(ChrInfo(seqname, len));
vector<int>& vec = 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[i]); }
}
for (int i = 1; i <= M; i++) {
if (seqs[i] == "") {
const Transcript& transcript = transcripts.getTranscriptAt(i);
fprintf(stderr, "Cannot extract transcript %s's sequence from chromosome %s! Loading chromosome %s's sequence is failed. Please check if 1) the chromosome directory is set correctly; 2) the list of chromosome files is complete; 3) the FASTA files containing chromosome sequences are not truncated or having wrong format.\n", \
transcript.getTranscriptID().c_str(), transcript.getSeqName().c_str(), transcript.getSeqName().c_str());
exit(-1);
}
}
sort(chrvec.begin(), chrvec.end());
if (verbose) { printf("Extracting sequences is done!\n"); }
writeResults(argv[1]);
return 0;
}
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"); }
}
int main(int argc, char* argv[]) {
ifstream fin;
bool quiet = false;
if (argc < 6) {
printf("Usage : rsem-run-em refName read_type sampleName imdName statName [-p #Threads] [-b samInpType samInpF has_fn_list_? [fn_list]] [-q] [--gibbs-out] [--sampling] [--seed seed] [--calc-evaluation-score nb_r nb_p L w]\n\n");
printf(" refName: reference name\n");
printf(" read_type: 0 single read without quality score; 1 single read with quality score; 2 paired-end read without quality score; 3 paired-end read with quality score.\n");
printf(" sampleName: sample's name, including the path\n");
printf(" sampleToken: sampleName excludes the path\n");
printf(" -p: number of threads which user wants to use. (default: 1)\n");
printf(" -b: produce bam format output file. (default: off)\n");
printf(" -q: set it quiet\n");
printf(" --gibbs-out: generate output file use by Gibbs sampler. (default: off)\n");
printf(" --sampling: sample each read from its posterior distribution when bam file is generated. (default: off)\n");
printf(" --seed uint32: the seed used for the BAM sampling. (default: off)\n");
printf(" --calc-evaluation-score nb_r nb_p L w: "
"nb_r and nb_p are parameters for the true transcript length distribution, which is modeled by a negative binomial distribution; "
"L is the read length and w is the mininum overlap required for joining two contigs.\n");
printf("// model parameters should be in imdName.mparams.\n");
exit(-1);
}
time_t a = time(NULL);
strcpy(refName, argv[1]);
read_type = atoi(argv[2]);
strcpy(outName, argv[3]);
strcpy(imdName, argv[4]);
strcpy(statName, argv[5]);
nThreads = 1;
genBamF = false;
bamSampling = false;
genGibbsOut = false;
calcEvalScore = false;
pt_fn_list = NULL;
hasSeed = false;
for (int i = 6; i < argc; i++) {
if (!strcmp(argv[i], "-p")) { nThreads = atoi(argv[i + 1]); }
if (!strcmp(argv[i], "-b")) {
genBamF = true;
inpSamType = argv[i + 1][0];
strcpy(inpSamF, argv[i + 2]);
if (atoi(argv[i + 3]) == 1) {
strcpy(fn_list, argv[i + 4]);
pt_fn_list = (char*)(&fn_list);
}
}
if (!strcmp(argv[i], "-q")) { quiet = true; }
if (!strcmp(argv[i], "--gibbs-out")) { genGibbsOut = true; }
if (!strcmp(argv[i], "--sampling")) { bamSampling = true; }
if (!strcmp(argv[i], "--seed")) {
hasSeed = true;
int len = strlen(argv[i + 1]);
seed = 0;
for (int k = 0; k < len; k++) seed = seed * 10 + (argv[i + 1][k] - '0');
}
if (!strcmp(argv[i], "--calc-evaluation-score")) {
calcEvalScore = true;
nb_r = atof(argv[i + 1]);
nb_p = atof(argv[i + 2]);
L = atoi(argv[i + 3]);
w = atoi(argv[i + 4]);
}
}
general_assert(nThreads > 0, "Number of threads should be bigger than 0!");
verbose = !quiet;
//basic info loading
sprintf(refF, "%s.seq", refName);
refs.loadRefs(refF);
M = refs.getM();
sprintf(tiF, "%s.ti", refName);
transcripts.readFrom(tiF);
sprintf(cntF, "%s.cnt", statName);
fin.open(cntF);
general_assert(fin.is_open(), "Cannot open " + cstrtos(cntF) + "! It may not exist.");
fin>>N0>>N1>>N2>>N_tot;
fin.close();
general_assert(N1 > 0, "There are no alignable reads!");
if ((READ_INT_TYPE)nThreads > N1) nThreads = N1;
//set model parameters
mparams.M = M;
mparams.N[0] = N0; mparams.N[1] = N1; mparams.N[2] = N2;
mparams.refs = &refs;
sprintf(mparamsF, "%s.mparams", imdName);
fin.open(mparamsF);
general_assert(fin.is_open(), "Cannot open " + cstrtos(mparamsF) + "It may not exist.");
fin>> mparams.minL>> mparams.maxL>> mparams.probF;
int val; // 0 or 1 , for estRSPD
fin>>val;
mparams.estRSPD = (val != 0);
fin>> mparams.B>> mparams.mate_minL>> mparams.mate_maxL>> mparams.mean>> mparams.sd;
fin>> mparams.seedLen;
fin.close();
//run EM
switch(read_type) {
case 0 : EM<SingleRead, SingleHit, SingleModel>(); break;
case 1 : EM<SingleReadQ, SingleHit, SingleQModel>(); break;
case 2 : EM<PairedEndRead, PairedEndHit, PairedEndModel>(); break;
case 3 : EM<PairedEndReadQ, PairedEndHit, PairedEndQModel>(); break;
default : fprintf(stderr, "Unknown Read Type!\n"); exit(-1);
}
if (calcEvalScore) {
CalcEvalScore ces(refs, nb_r, nb_p, L, w, statName);
sprintf(scoreF, "%s.score", outName);
ces.writeScoresTo(scoreF);
char groupF[STRLEN];
GroupInfo gi;
sprintf(groupF, "%s.grp", argv[1]);
gi.load(groupF);
ces.generateExpressionFiles(gi, transcripts, scoreF);
}
time_t b = time(NULL);
printTimeUsed(a, b, "EM.cpp");
return 0;
}
void writeResults(int option, char* refName) {
ofstream fout, fout2;
string cur_gene_id, cur_transcript_id, name;
vector<int> gi, gt, ta;
sprintf(tiF, "%s.ti", refName);
transcripts.writeTo(tiF);
if (verbose) { printf("Transcript Information File is generated!\n"); }
cur_gene_id = ""; gi.clear();
if (option == 2) { cur_transcript_id = ""; gt.clear(); ta.clear(); }
for (int i = 1; i <= M; i++) {
const Transcript& transcript = transcripts.getTranscriptAt(i);
if (cur_gene_id != transcript.getGeneID()) {
gi.push_back(i);
if (option == 2) gt.push_back((int)ta.size());
cur_gene_id = transcript.getGeneID();
}
if ((option == 2) && (cur_transcript_id != transcript.getTranscriptID())) {
ta.push_back(i);
cur_transcript_id = transcript.getTranscriptID();
}
}
gi.push_back(M + 1);
if (option == 2) { gt.push_back((int)ta.size()); ta.push_back(M + 1); }
sprintf(groupF, "%s.grp", refName);
fout.open(groupF);
for (int i = 0; i < (int)gi.size(); i++) fout<< gi[i]<< endl;
fout.close();
if (verbose) { printf("Group File is generated!\n"); }
if (option == 2) {
sprintf(gtF, "%s.gt", refName);
fout.open(gtF);
for (int i = 0; i < (int)gt.size(); i++) fout<< gt[i]<< endl;
fout.close();
sprintf(taF, "%s.ta", refName);
fout.open(taF);
for (int i = 0; i < (int)ta.size(); i++) fout<< ta[i]<< endl;
fout.close();
if (verbose) { printf("Allele-specific group files are generated!\n"); }
}
sprintf(refFastaF, "%s.transcripts.fa", refName);
sprintf(chromListF, "%s.chrlist", refName);
fout2.open(chromListF);
fout.open(refFastaF);
for (int i = 1; i <= M; i++) {
name = transcripts.getTranscriptAt(i).getSeqName();
iter = name2seq.find(name);
general_assert(iter != name2seq.end(), "Cannot recognize sequence ID" + name + "!");
fout<<">"<<name<<endl;
fout<<iter->second<<endl;
fout2<<name<<'\t'<<iter->second.length()<<endl;
}
fout.close();
fout2.close();
if (verbose) {
printf("Chromosome List File is generated!\n");
printf("Extracted Sequences File is generated!\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;
}
