AlnBlock * Align_Sequences_ProteinBlockAligner(Sequence * one,Sequence * two,CompMat * comp,int bentry,int bexit,int b1exit,int b_self,int b_on,DPRunImpl * dpri)
{
ComplexSequence * cone=NULL;
ComplexSequence * ctwo=NULL;
AlnBlock * alb= NULL;
PackAln * pal=NULL;
ComplexSequenceEvalSet * evalfunc = NULL;
if( one == NULL || two == NULL || comp == NULL ) {
warn("Passed in NULL objects into Align_Sequences_ProteinSmithWaterman!");
return NULL;
}
if( one->type != SEQUENCE_PROTEIN ) {
warn("Sequence %s is not typed as protein... ignoring!\n",one->name);
}
if( two->type != SEQUENCE_PROTEIN ) {
warn("Sequence %s is not typed as protein... ignoring!\n",two->name);
}
evalfunc = default_aminoacid_ComplexSequenceEvalSet();
cone = new_ComplexSequence(one,evalfunc);
if( cone == NULL )
goto cleanup;
ctwo = new_ComplexSequence(two,evalfunc);
if( ctwo == NULL )
goto cleanup;
pal = PackAln_bestmemory_ProteinBlockAligner(cone,ctwo,comp,bentry,b1exit,b_on,b_self,bexit,NULL,dpri);
if( pal == NULL )
goto cleanup;
alb = convert_PackAln_to_AlnBlock_ProteinSW(pal);
goto cleanup;
cleanup :
if( cone != NULL )
free_ComplexSequence(cone);
if( ctwo != NULL )
free_ComplexSequence(ctwo);
if( pal != NULL )
free_PackAln(pal);
if( evalfunc != NULL )
free_ComplexSequenceEvalSet(evalfunc);
return alb;
}
boolean free_temporary_objects(void)
{
alb = free_AlnBlock(alb);
pal = free_PackAln(pal);
mss = free_MatchSummarySet(mss);
gr = free_GenomicRegion(gr);
return TRUE;
}
DnaProfileMatchPair * DnaProfileMatchPair_from_DnaProfile(DnaProfile * query,DnaProfile * target,DnaProfileEnginePara * dpep)
{
DnaProfileMatchPair * out;
DnaProfileScore * query_s;
DnaProfileScore * target_s;
DnaProfileMatchScore * match;
PackAln * pal;
assert(query != NULL);
assert(target != NULL);
/*
assert(query->len > 4 );
assert(target->len > 4);
*/
out = DnaProfileMatchPair_alloc();
out->query = hard_link_DnaProfile(query);
out->target = hard_link_DnaProfile(target);
query_s = DnaProfileScore_from_DnaProfile(query);
target_s = DnaProfileScore_from_DnaProfile(target);
fprintf(stderr,"Matching %d to %d\n",query->len,target->len);
match= new_ALLR_DnaProfileMatchScore(query,target);
pal = PackAln_bestmemory_DnaProfileMat(query_s,target_s,match,Probability2Score(dpep->open_unmatched),Probability2Score(dpep->ext_unmatched),Probability2Score(dpep->gap_unmatched),NULL,dpep->dpri);
fprintf(stderr,"...Made pal %d\n",pal);
out->alb = convert_PackAln_to_AlnBlock_DnaProfileMat(pal);
out->score = pal->score;
fprintf(stderr,"...freeing pal\n");
free_PackAln(pal);
fprintf(stderr,"...freeing match\n");
free_DnaProfileMatchScore(match);
fprintf(stderr,"...freeing query\n");
free_DnaProfileScore(query_s);
fprintf(stderr,"...freeing target\n");
free_DnaProfileScore(target_s);
return out;
}
AlnBlock * Align_Sequences_ProteinSmithWaterman(Sequence * one,Sequence * two,CompMat * comp,int gap,int ext,DPRunImpl * dpri)
{
AlnBlock * out = NULL;
ComplexSequenceEvalSet * evalfunc = NULL;
ComplexSequence * query_cs = NULL;
ComplexSequence * target_cs = NULL;
PackAln * pal = NULL;
if( one == NULL || two == NULL || comp == NULL ) {
warn("Passed in NULL objects into Align_Sequences_ProteinSmithWaterman!");
return NULL;
}
if( one->type != SEQUENCE_PROTEIN ) {
warn("Sequence %s is not typed as protein... ignoring!\n",one->name);
}
if( two->type != SEQUENCE_PROTEIN ) {
warn("Sequence %s is not typed as protein... ignoring!\n",two->name);
}
if( gap > 0 || ext > 0 ) {
warn("Gap penalties %d,%d only make sense if they are negative",gap,ext);
return NULL;
}
evalfunc = default_aminoacid_ComplexSequenceEvalSet();
query_cs = new_ComplexSequence(one,evalfunc);
if( query_cs == NULL )
goto cleanup;
target_cs = new_ComplexSequence(two,evalfunc);
if( target_cs == NULL )
goto cleanup;
pal = PackAln_bestmemory_ProteinSW(query_cs,target_cs,comp,gap,ext,NULL,dpri);
if( pal == NULL )
goto cleanup;
out = convert_PackAln_to_AlnBlock_ProteinSW(pal);
goto cleanup;
cleanup :
if( query_cs != NULL )
free_ComplexSequence(query_cs);
if( target_cs != NULL )
free_ComplexSequence(target_cs);
if( pal != NULL )
free_PackAln(pal);
if( evalfunc != NULL )
free_ComplexSequenceEvalSet(evalfunc);
return out;
}
boolean show_output(void)
{
int i,k;
ThreeStateModel * temptsm;
AlnBlock * alb;
PackAln * pal;
MatchSummarySet * mss;
Protein * ps;
cDNA * cdna;
double bits;
boolean fitted_res = FALSE;
AlnBlockList * alist;
AlnBlock * anchored;
SequenceSet * set;
AlnColumn * alt;
Protein * trans;
/* sort by bit score first */
sort_Hscore_by_score(hs);
if( search_mode == PC_SEARCH_S2DB ) {
if( hs->his == NULL || hs->his->total < 1000 ) {
info("Cannot fit histogram to a db smaller than 1,000");
fprintf(ofp,"[Warning: Can't fit histogram to a db smaller than 1,000]\n\n");
show_histogram = FALSE;
} else {
fitted_res = TRUE;
fit_Hscore_to_EVD(hs,20);
}
}
/* deal with initialising anchored alignment.
* Could be done for either single HMMs or single proteins,
* but we will only do it for HMMs at the moment
*/
if( make_anchored_aln == TRUE ) {
if( tsm == NULL ) {
warn("Attempting to make an achored alignment without a HMM. impossible!");
make_anchored_aln = FALSE;
} else {
anchored = single_unit_AlnBlock(tsm->len,"MATCH_STATE");
set = SequenceSet_alloc_std();
}
}
/* dofus catcher */
if( aln_alg != alg ) {
fprintf(ofp,"\n#\n#WARNING!\n#\n# Your alignment algorithm is different from your search algorithm.\n# This is probably quite sensible but will lead to differing scores.\n# Use the search score as an indicator of the significance of the match\n# Read the docs for more information\n#\n");
}
fprintf(ofp,"\n\n#High Score list\n");
fprintf(ofp,"#Protein ID DNA Str ID Bits Evalue\n");
fprintf(ofp,"--------------------------------------------------------------------------\n");
for(i=0;i<hs->len;i++) {
bits = Score2Bits(hs->ds[i]->score);
if( bits < search_cutoff ) {
break;
}
if( fitted_res == TRUE && evalue_search_str != NULL ) {
if( hs->ds[i]->evalue > evalue_search_cutoff )
break;
}
if( fitted_res == TRUE)
fprintf(ofp,"Protein %-20sDNA [%c] %-24s %.2f %.2g\n",hs->ds[i]->query->name,hs->ds[i]->target->is_reversed == TRUE ? '-' : '+',hs->ds[i]->target->name,bits,hs->ds[i]->evalue);
else
fprintf(ofp,"Protein %-20sDNA [%c] %-24s %.2f\n",hs->ds[i]->query->name,hs->ds[i]->target->is_reversed == TRUE ? '-' : '+',hs->ds[i]->target->name,bits);
}
if( search_mode == PC_SEARCH_S2DB && show_histogram == TRUE ) {
fprintf(ofp,"\n\n#Histogram\n");
fprintf(ofp,"-----------------------------------------------------------------------\n");
PrintASCIIHistogram(hs->his,ofp);
}
fprintf(ofp,"\n\n#Alignments\n");
fprintf(ofp,"-----------------------------------------------------------------------\n");
for(i=0;i<hs->len;i++) {
bits = Score2Bits(hs->ds[i]->score);
if( bits < search_cutoff ) {
break;
}
if( i >= aln_number ) {
break;
}
if( fitted_res == TRUE && evalue_search_str != NULL ) {
if( hs->ds[i]->evalue > evalue_search_cutoff )
break;
}
fprintf(ofp,"\n\n>Results for %s vs %s (%s) [%d]\n",hs->ds[i]->query->name,hs->ds[i]->target->name,hs->ds[i]->target->is_reversed == TRUE ? "reverse" : "forward",i+1 );
cdna = get_cDNA_from_cDNADB(cdb,hs->ds[i]->target);
temptsm = indexed_ThreeStateModel_ThreeStateDB(tsmdb,hs->ds[i]->query);
alb = AlnBlock_from_TSM_estwise_wrap(temptsm,cdna,cps,cm,ct,rmd,aln_alg,use_syn,allN,flat_insert,dpri,&pal);
if( alb == NULL ) {
warn("Got a NULL alignment. Exiting now due to presumed problems");
fprintf(ofp,"\n\n*Got a NULL alignment. Exiting now due to presumed problems*\n\n");
return FALSE;
}
if( use_single_pro == FALSE)
mss = MatchSummarySet_from_AlnBlock_genewise(alb,temptsm->name,1,cdna->baseseq);
else
mss = MatchSummarySet_from_AlnBlock_genewise(alb,pro->baseseq->name,pro->baseseq->offset,cdna->baseseq);
if( show_pretty == TRUE ) {
fprintf(ofp,"\n%s output\nScore %4.2f bits over entire alignment.\nThis will be different from per-alignment scores. See manual for details\nFor computer parsable output, try %s -help or read the manual\n",program_name,Score2Bits(pal->score),program_name);
if( use_syn == FALSE ) {
fprintf(ofp,"Scores as bits over a flat simple random model\n\n");
} else {
fprintf(ofp,"Scores as bits over a synchronous coding model\n\n");
}
ps = pseudo_Protein_from_ThreeStateModel(temptsm);
protcdna_ascii_display(alb,ps->baseseq->seq,ps->baseseq->name,ps->baseseq->offset,cdna,ct,15,main_block,TRUE,ofp);
free_Protein(ps);
fprintf(ofp,"%s\n",divide_str);
}
if( show_match_sum == TRUE ) {
show_MatchSummary_genewise_header(ofp);
show_MatchSummarySet_genewise(mss,ofp);
fprintf(ofp,"%s\n",divide_str);
}
if( show_pep == TRUE ) {
alt = alb->start;
for(;alt != NULL;) {
trans = Protein_from_GeneWise_AlnColumn(cdna->baseseq,alt,1,&alt,ct,is_random_AlnColumn_genewise);
if ( trans == NULL )
break;
write_fasta_Sequence(trans->baseseq,ofp);
free_Protein(trans);
}
fprintf(ofp,"%s\n",divide_str);
}
if( show_AlnBlock == TRUE ) {
mapped_ascii_AlnBlock(alb,Score2Bits,0,ofp);
fprintf(ofp,"%s\n",divide_str);
}
if( show_PackAln == TRUE ) {
show_simple_PackAln(pal,ofp);
fprintf(ofp,"%s\n",divide_str);
}
/*
* This goes at the end because it destroys the alb structure
*/
if( make_anchored_aln == TRUE ) {
/* attach sequence to als in alb, so we have it for later use */
alb->seq[1]->data = (void *) cdna->baseseq;
/* add to SequenceSet so we can destroy the memory */
add_SequenceSet(set,hard_link_Sequence(cdna->baseseq));
alist = split_AlnBlock(alb,is_random_AlnColumn_genewise);
for(k=0;k<alist->len;k++) {
/* actually produce the anchored alignment */
/*mapped_ascii_AlnBlock(alist->alb[k],Score2Bits,stderr);*/
add_to_anchored_AlnBlock(anchored,alist->alb[k]);
/* dump_ascii_AlnBlock(anchored,stderr);*/
}
}
alb = free_AlnBlock(alb);
pal = free_PackAln(pal);
mss = free_MatchSummarySet(mss);
cdna = free_cDNA(cdna);
temptsm = free_ThreeStateModel(temptsm);
}
if( do_complete_analysis == TRUE ) {
fprintf(ofp,"\n\n#Complete Analysis\n");
fprintf(ofp,"-------------------------------------------------------------\n\n");
/* ok - end of loop over relevant hits. If we have an
* anchored alignment, print it out!
*/
if( make_anchored_aln == TRUE ) {
/*dump_ascii_AlnBlock(anchored,stderr);*/
write_mul_estwise_AlnBlock(anchored,ct,ofp);
fprintf(ofp,"%s\n",divide_str);
}
}
return TRUE;
}
int main(int argc,char ** argv)
{
Sequence * query;
Sequence * target;
ComplexSequence * query_cs;
ComplexSequence * target_cs;
ComplexSequenceEvalSet * evalfunc;
CompMat * comp;
char * comp_file;
int gap = (12);
int ext = (2);
boolean show_raw_output = FALSE;
boolean show_label_output = FALSE;
boolean show_fancy_output = FALSE;
boolean has_outputted = FALSE;
PackAln * pal;
AlnBlock * alb;
/*
* Process command line options
* -h or -help gives us help
* -g for gap value (an int) - rely on commandline error processing
* -e for ext value (an int) - rely on commandline error processing
* -m for matrix (a char)
* -r - raw matrix output
* -l - label output
* -f - fancy output
*
*
* Use calls to commandline.h functions
*
*/
if( strip_out_boolean_argument(&argc,argv,"h") == TRUE || strip_out_boolean_argument(&argc,argv,"-help") == TRUE) {
show_help(stdout);
exit(1);
}
show_raw_output = strip_out_boolean_argument(&argc,argv,"r");
show_label_output = strip_out_boolean_argument(&argc,argv,"l");
show_fancy_output = strip_out_boolean_argument(&argc,argv,"f");
/** if all FALSE, set fancy to TRUE **/
if( show_raw_output == FALSE && show_label_output == FALSE )
show_fancy_output = TRUE;
(void) strip_out_integer_argument(&argc,argv,"g",&gap);
(void) strip_out_integer_argument(&argc,argv,"e",&ext);
comp_file = strip_out_assigned_argument(&argc,argv,"m");
if( comp_file == NULL)
comp_file = "blosum62.bla";
if( argc != 3 ) {
warn("Must have two arguments for sequence 1 and sequence 2 %d",argc);
show_help(stdout);
exit(1);
}
/*
* Read in two sequences
*/
if( (query=read_fasta_file_Sequence(argv[1])) == NULL ) {
fatal("Unable to read the sequence in file %s",argv[1]);
}
if( (target=read_fasta_file_Sequence(argv[2])) == NULL ) {
fatal("Unable to read the sequence in file %s",argv[2]);
}
/*
* Open a blosum matrix. This will be opened from WISECONFIGDIR
* or WISEPERSONALDIR if it is not present in the current directory.
*/
comp = read_Blast_file_CompMat(comp_file);
if( comp == NULL ) {
fatal("unable to read file %s",comp_file);
}
/*
* Convert sequences to ComplexSequences:
* To do this we need an protein ComplexSequenceEvalSet
*
*/
evalfunc = default_aminoacid_ComplexSequenceEvalSet();
query_cs = new_ComplexSequence(query,evalfunc);
if( query_cs == NULL ) {
fatal("Unable to make a protein complex sequence from %s",query->name);
}
target_cs = new_ComplexSequence(target,evalfunc);
if( target_cs == NULL ) {
fatal("Unable to make a protein complex sequence from %s",target->name);
}
/*
* Make an alignment. I don't care about the implementation:
* If the sequences are small enough then it should use explicit memory.
* Long sequences should use divide and conquor methods.
*
* Calling PackAln_bestmemory_ProteinSW is the answer
* This function decides on the best method considering the
* memory and changes accordingly. It frees the matrix memory
* at the end as well.
*
*/
pal = PackAln_bestmemory_ProteinSW(query_cs,target_cs,comp,-gap,-ext,NULL);
if( pal == NULL ) {
fatal("Unable to make an alignment from %s and %s",query->name,target->name);
}
/*
* ok, make other alignment forms, and be ready to show
*/
alb = convert_PackAln_to_AlnBlock_ProteinSW(pal);
/*
* show output. If multiple outputs, divide using //
*/
if( show_raw_output == TRUE ) {
show_simple_PackAln(pal,stdout);
puts("//\n");
}
if( show_label_output == TRUE ) {
show_flat_AlnBlock(alb,stdout);
}
if( show_fancy_output == TRUE ) {
write_pretty_seq_align(alb,query,target,15,50,stdout);
puts("//\n");
}
/*
* Destroy the memory.
*/
free_Sequence(query);
free_Sequence(target);
free_CompMat(comp);
free_ComplexSequence(query_cs);
free_ComplexSequence(target_cs);
free_PackAln(pal);
free_AlnBlock(alb);
return 0;
}
