int main(int argc, const char **argv) {
Parameters par(argc, argv);
strcpy(par.infile, "");
strcpy(par.outfile, "");
// maximum number of sequences to be written
par.nseqdis = par.maxseq - 1;
// no filtering for maximum diversity
par.Ndiff = 0;
ProcessArguments(par);
// Check command line input and default values
if (!*par.infile) {
help(par);
HH_LOG(ERROR) << "Input file is missing!" << std::endl;
exit(4);
}
if (!*par.outfile) {
help(par);
HH_LOG(ERROR) << "Output file is missing!" << std::endl;
exit(4);
}
HH_LOG(INFO) << "Input file = " << par.infile << "\n";
HH_LOG(INFO) << "Output file = " << par.outfile << "\n";
// Reads in an alignment from par.infile into matrix X[k][l] as ASCII
FILE* inf = NULL;
if (strcmp(par.infile, "stdin")) {
inf = fopen(par.infile, "r");
if (!inf) {
OpenFileError(par.infile, __FILE__, __LINE__, __func__);
}
}
else {
inf = stdin;
}
Alignment qali(par.maxseq, par.maxres);
qali.Read(inf, par.infile, par.mark, par.maxcol, par.nseqdis);
fclose(inf);
// Convert ASCII to int (0-20),throw out all insert states, record their number in I[k][i]
// and store marked sequences in name[k] and seq[k]
qali.Compress(par.infile, par.cons, par.maxcol, par.M, par.Mgaps);
// substitution matrix flavours
float __attribute__((aligned(16))) P[20][20];
float __attribute__((aligned(16))) R[20][20];
float __attribute__((aligned(16))) Sim[20][20];
float __attribute__((aligned(16))) S[20][20];
float __attribute__((aligned(16))) pb[21];
SetSubstitutionMatrix(par.matrix, pb, P, R, S, Sim);
// Remove sequences with seq. identity larger than seqid percent (remove the shorter of two)
qali.N_filtered = qali.Filter(par.max_seqid, S, par.coverage, par.qid, par.qsc,par.Ndiff);
// Atune alignment diversity q.Neff with qsc to value Neff_goal
if (par.Neff >= 1.0) {
qali.FilterNeff(par.wg, par.mark, par.cons, par.showcons, par.max_seqid, par.coverage, par.Neff, pb, S, Sim);
}
// Write filtered alignment WITH insert states (lower case) to alignment file
qali.WriteToFile(par.outfile, par.append);
}
/////////////////////////////////////////////////////////////////////////////////////
//// MAIN PROGRAM
/////////////////////////////////////////////////////////////////////////////////////
int main(int argc, char **argv) {
char* argv_conf[MAXOPT]; // Input arguments from .hhdefaults file (first=1: argv_conf[0] is not used)
int argc_conf; // Number of arguments in argv_conf
strcpy(par.infile, "");
strcpy(par.outfile, "");
strcpy(par.alnfile, "");
//Default parameter settings
par.nseqdis = MAXSEQ - 1; // maximum number of sequences to be written
par.showcons = 0;
par.cons = 1;
par.Ndiff = 0;
par.max_seqid = 100;
par.coverage = 0;
par.pc_hhm_context_engine.pca = 0.0; // no amino acid pseudocounts
par.pc_hhm_nocontext_a = 0.0; // no amino acid pseudocounts
par.gapb = 0.0; // no transition pseudocounts
// Make command line input globally available
par.argv = argv;
par.argc = argc;
RemovePathAndExtension(program_name, argv[0]);
// Enable changing verbose mode before defaults file and command line are processed
int v = 2;
for (int i = 1; i < argc; i++) {
if (!strcmp(argv[i], "-def"))
par.readdefaultsfile = 1;
else if (strcmp(argv[i], "-v") == 0) {
v = atoi(argv[i + 1]);
}
}
par.v = Log::from_int(v);
Log::reporting_level() = par.v;
par.SetDefaultPaths();
// Read .hhdefaults file?
if (par.readdefaultsfile) {
// Process default otpions from .hhconfig file
ReadDefaultsFile(argc_conf, argv_conf);
ProcessArguments(argc_conf, argv_conf);
}
// Process command line options (they override defaults from .hhdefaults file)
ProcessArguments(argc, argv);
Alignment* qali = new Alignment(MAXSEQ, par.maxres);
HMM* q = new HMM(MAXSEQDIS, par.maxres); //Create a HMM with maximum of par.maxres match states
// q is only available after maxres is known, so we had to move this here
for (int i = 1; i <= argc - 1; i++) {
if (!strcmp(argv[i], "-name") && (i < argc - 1)) {
strmcpy(q->name, argv[++i], NAMELEN - 1); //copy longname to name...
strmcpy(q->longname, argv[i], DESCLEN - 1); //copy full name to longname
}
}
// Check command line input and default values
if (!*par.infile) {
help();
HH_LOG(ERROR) << "Input file is missing!" << std::endl;
exit(4);
}
// Get basename
RemoveExtension(q->file, par.infile); //Get basename of infile (w/o extension):
// Outfile not given? Name it basename.hhm
if (!*par.outfile && !*par.alnfile) {
RemoveExtension(par.outfile, par.infile);
strcat(par.outfile, ".seq");
}
// Prepare CS pseudocounts lib
if (!par.nocontxt && *par.clusterfile) {
InitializePseudocountsEngine(par, context_lib, crf, pc_hhm_context_engine,
pc_hhm_context_mode, pc_prefilter_context_engine,
pc_prefilter_context_mode);
}
// Set substitution matrix; adjust to query aa distribution if par.pcm==3
SetSubstitutionMatrix(par.matrix, pb, P, R, S, Sim);
// Read input file (HMM, HHM, or alignment format), and add pseudocounts etc.
char input_format = 0;
ReadQueryFile(par, par.infile, input_format, par.wg, q, qali, pb, S, Sim);
// Same code as in PrepareQueryHMM(par.infile,input_format,q,qali), except that we add SS prediction
// Add Pseudocounts, if no HMMER input
if (input_format == 0) {
// Transform transition freqs to lin space if not already done
q->AddTransitionPseudocounts(par.gapd, par.gape, par.gapf, par.gapg,
par.gaph, par.gapi, par.gapb, par.gapb);
// Comput substitution matrix pseudocounts
if (par.nocontxt) {
// Generate an amino acid frequency matrix from f[i][a] with full pseudocount admixture (tau=1) -> g[i][a]
q->PreparePseudocounts(R);
// Add amino acid pseudocounts to query: p[i][a] = (1-tau)*f[i][a] + tau*g[i][a]
q->AddAminoAcidPseudocounts(par.pc_hhm_nocontext_mode,
par.pc_hhm_nocontext_a, par.pc_hhm_nocontext_b,
par.pc_hhm_nocontext_c);
}
else {
// Add full context specific pseudocounts to query
q->AddContextSpecificPseudocounts(pc_hhm_context_engine,
pc_hhm_context_mode);
}
}
else {
q->AddAminoAcidPseudocounts(0, par.pc_hhm_nocontext_a,
par.pc_hhm_nocontext_b, par.pc_hhm_nocontext_c);
}
q->CalculateAminoAcidBackground(pb);
if (par.columnscore == 5 && !q->divided_by_local_bg_freqs)
q->DivideBySqrtOfLocalBackgroundFreqs(
par.half_window_size_local_aa_bg_freqs, pb);
// Write consensus sequence to sequence file
// Consensus sequence is calculated in hhalignment.C, Alignment::FrequenciesAndTransitions()
if (*par.outfile) {
FILE* outf = NULL;
if (strcmp(par.outfile, "stdout")) {
outf = fopen(par.outfile, "a");
if (!outf)
OpenFileError(par.outfile, __FILE__, __LINE__, __func__);
}
else
outf = stdout;
// OLD
//// ">name_consensus" -> ">name consensus"
//strsubst(q->sname[q->nfirst],"_consensus"," consensus");
//fprintf(outf,">%s\n%s\n",q->sname[q->nfirst],q->seq[q->nfirst]+1);
// NEW (long header needed for NR30cons database)
fprintf(outf, ">%s\n%s\n", q->longname, q->seq[q->nfirst] + 1);
fclose(outf);
}
// Print A3M/A2M/FASTA output alignment
if (*par.alnfile) {
HalfAlignment qa;
int n = imin(q->n_display,
par.nseqdis + (q->nss_dssp >= 0) + (q->nss_pred >= 0)
+ (q->nss_conf >= 0) + (q->ncons >= 0));
qa.Set(q->name, q->seq, q->sname, n, q->L, q->nss_dssp, q->nss_pred,
q->nss_conf, q->nsa_dssp, q->ncons);
if (par.outformat == 1)
qa.BuildFASTA();
else if (par.outformat == 2)
qa.BuildA2M();
else if (par.outformat == 3)
qa.BuildA3M();
if (qali->readCommentLine)
qa.Print(par.alnfile, par.append, qali->longname); // print alignment to outfile
else
qa.Print(par.alnfile, par.append); // print alignment to outfile
}
delete qali;
delete q;
DeletePseudocountsEngine(context_lib, crf, pc_hhm_context_engine,
pc_hhm_context_mode, pc_prefilter_context_engine,
pc_prefilter_context_mode);
}
