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;
}
int main(int argc, char* argv[]) {
if (argc < 8) {
printf("Usage: rsem-calculate-credibility-intervals reference_name sample_name sampleToken confidence nCV nSpC nMB [-p #Threads] [-q]\n");
exit(-1);
}
confidence = atof(argv[4]);
nCV = atoi(argv[5]);
nSpC = atoi(argv[6]);
nMB = atoi(argv[7]);
nThreads = 1;
quiet = false;
for (int i = 8; i < argc; i++) {
if (!strcmp(argv[i], "-p")) nThreads = atoi(argv[i + 1]);
if (!strcmp(argv[i], "-q")) quiet = true;
}
verbose = !quiet;
sprintf(refF, "%s.seq", argv[1]);
refs.loadRefs(refF, 1);
M = refs.getM();
sprintf(groupF, "%s.grp", argv[1]);
gi.load(groupF);
m = gi.getm();
nSamples = nCV * nSpC;
cvlen = M + 1;
assert(nSamples > 0 && cvlen > 1); // for Buffter.h: (bufsize_type)nSamples
sprintf(imdName, "%s.temp/%s", argv[2], argv[3]);
sprintf(statName, "%s.stat/%s", argv[2], argv[3]);
sprintf(tmpF, "%s.tmp", imdName);
sprintf(cvsF, "%s.countvectors", imdName);
sprintf(modelF, "%s.model", statName);
FILE *fi = fopen(modelF, "r");
general_assert(fi != NULL, "Cannot open " + cstrtos(modelF) + "!");
assert(fscanf(fi, "%d", &model_type) == 1);
fclose(fi);
// Phase I
switch(model_type) {
case 0 : sample_theta_vectors_from_count_vectors<SingleModel>(); break;
case 1 : sample_theta_vectors_from_count_vectors<SingleQModel>(); break;
case 2 : sample_theta_vectors_from_count_vectors<PairedEndModel>(); break;
case 3 : sample_theta_vectors_from_count_vectors<PairedEndQModel>(); break;
}
// Phase II
calculate_credibility_intervals(imdName);
/*
sprintf(command, "rm -f %s", tmpF);
int status = system(command);
if (status != 0) {
fprintf(stderr, "Cannot delete %s!\n", tmpF);
exit(-1);
}
*/
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 {
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;
}
