boolean show_pretty_aln(void)
{
Protein * ps;
fprintf(ofp,"\n%s output\nScore %4.2f bits over entire alignment\n",program_name,Score2Bits(pal->score));
if( alg == GWWRAP_2193L || alg == GWWRAP_2193) {
fprintf(ofp,"Entrie alignment score contains unseen 'random' score segments\nYou should only use the per-alignments score printed below\nfor the bits score of the alignment\n\n");
}
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");
}
if( use_tsm == FALSE ) {
fprintf(ofp,"Warning: The bits scores is not probablistically correct for single seqs\nSee WWW help for more info\n\n");
protgene_ascii_display(alb,pro->baseseq->seq,pro->baseseq->name,pro->baseseq->offset,gen,ct,15,main_block,(alg == GWWRAP_623L || alg == GWWRAP_2193L || alg == GWWRAP_2193) ? TRUE : FALSE, ofp);
} else {
ps = pseudo_Protein_from_ThreeStateModel(tsm);
protgene_ascii_display(alb,ps->baseseq->seq,ps->baseseq->name,ps->baseseq->offset,gen,ct,15,main_block,(alg == GWWRAP_623L || alg == GWWRAP_2193L || alg == GWWRAP_2193) ? TRUE : FALSE,ofp);
free_Protein(ps);
}
fprintf(ofp,"%s\n",divide_str);
return TRUE;
}
int main(int argc,char ** argv)
{
int i;
SequenceSet * in;
Sequence * trans;
ThreeStateDB * tsd;
DPRunImpl * dpri;
CodonTable * ct;
int return_status;
ThreeStateModel * tsm;
ThreeStateScore * tss;
Protein * hmmp;
ComplexSequence * cs;
ComplexSequenceEvalSet * cses;
PackAln * pal;
AlnBlock * alb;
int show_align = 0;
int show_alb = 0;
int show_verbose = 1;
int show_trans = 0;
ct = read_CodonTable_file("codon.table");
cses = default_aminoacid_ComplexSequenceEvalSet();
dpri = new_DPRunImpl_from_argv(&argc,argv);
strip_out_boolean_def_argument(&argc,argv,"pretty",&show_align);
strip_out_boolean_def_argument(&argc,argv,"alb",&show_alb);
strip_out_boolean_def_argument(&argc,argv,"trans",&show_trans);
if( argc != 3 ) {
show_help(stdout);
exit(63);
}
in = read_fasta_SequenceSet_file(argv[1]);
tsd = HMMer2_ThreeStateDB(argv[2]);
assert(in);
assert(tsd);
assert(in->len == 2);
trans = translate_Sequence(in->set[0],ct);
if( show_trans ) {
write_fasta_Sequence(trans,stdout);
}
cs = new_ComplexSequence(trans,cses);
open_ThreeStateDB(tsd);
while( (tsm = read_TSM_ThreeStateDB(tsd,&return_status)) != NULL ) {
fold_RandomModel_into_ThreeStateModel(tsm,tsm->rm);
set_startend_policy_ThreeStateModel(tsm,TSM_local,10,1.0);
tss = ThreeStateScore_from_ThreeStateModel(tsm);
hmmp = pseudo_Protein_from_ThreeStateModel(tsm);
pal = PackAln_bestmemory_ThreeStateLoop(tss,cs,NULL,dpri);
alb = convert_PackAln_to_AlnBlock_ThreeStateLoop(pal);
if( show_alb ) {
show_flat_AlnBlock(alb,stdout);
}
if( show_align ) {
write_pretty_seq_align(alb,hmmp->baseseq,trans,15,50,stdout);
}
if( show_verbose ) {
show_verbose_evo(alb,tsm,in->set[0],in->set[1],ct,stdout);
}
}
}
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;
}
void show_verbose_evo(AlnBlock * alb,ThreeStateModel * tsm,Sequence * ref,Sequence * diff,CodonTable * ct,FILE * ofp)
{
AlnColumn * alc;
Protein * hmmp;
Sequence * ref_trans;
Sequence * diff_trans;
DnaProbMatrix * negative_dm;
DnaProbMatrix * pseudo_dm;
int i;
int count = 0;
double est_mutation = 0.0;
int dna_offset;
Score total_pseudo = 0;
Score total_neg = 0;
Score pseudo = 0;
Score neg = 0;
int count_ref_positive = 0;
int count_ref_negative = 0;
int count_ref_negative_0_5 = 0;
int count_ref_negative_5_10 = 0;
int count_ref_negative_10_15 = 0;
int syn_sites = 0;
int nonsyn_sites = 0;
int syn_changes = 0;
int nonsyn_changes = 0;
int diff_score;
char diff_aa;
char ref_aa;
int score_ratio = 0;
Score score_neg_5 = Probability2Score(Bits2Probability(-5.0));
Score score_neg_10 = Probability2Score(Bits2Probability(-10.0));
int k;
for(i=0;i<ref->len;i+=3) {
/* if this has changed, then it is definitely non syn */
if( aminoacid_from_seq(ct,ref->seq+i) != aminoacid_from_seq(ct,diff->seq+i)) {
for(k=0;k<3;k++) {
if( ref->seq[i+k] != diff->seq[i+k] ) {
nonsyn_changes++;
}
}
} else {
/* could still be syn change */
for(k=0;k<3;k++) {
if( ref->seq[i+k] != diff->seq[i+k] ) {
syn_changes++;
}
}
}
/* calculate the sites. There is always 2 non syn sites */
nonsyn_sites += 2;
if( four_fold_sites_CodonTable(ct,ref->seq+i) > 0 ) {
syn_sites++;
} else {
nonsyn_sites += 1;
}
}
for(i=0;i<ref->len;i++) {
if( ref->seq[i] != diff->seq[i] ) {
count++;
}
}
est_mutation = (double)count / (double)ref->len;
pseudo_dm = DnaProbMatrix_from_match(1.0 - est_mutation,NMaskType_BANNED);
negative_dm = DnaProbMatrix_from_match(1.0 - (est_mutation*2),NMaskType_BANNED);
ref_trans = translate_Sequence(ref,ct);
diff_trans = translate_Sequence(diff,ct);
hmmp = pseudo_Protein_from_ThreeStateModel(tsm);
for(alc=alb->start;alc != NULL;alc = alc->next) {
/* fprintf(stdout,"In position %s\n",alc->alu[0]->text_label); */
if( strcmp(alc->alu[0]->text_label,"SEQUENCE") == 0 &&
strcmp(alc->alu[1]->text_label,"SEQUENCE") == 0 ) {
dna_offset = alc->alu[1]->end*3;
pseudo = logl_pseudogene(ref->seq+dna_offset,diff->seq+dna_offset,pseudo_dm);
neg = logl_negative_selection(ref->seq+dna_offset,diff->seq+dna_offset,tsm->unit[alc->alu[0]->end],ct,
pseudo_dm);
/*
fprintf(ofp,"Position %d [%c], vs %d [%c,%c] Scores Negative %d, Pseudo %d\n",
alc->alu[0]->end,hmmp->baseseq->seq[alc->alu[0]->end],
alc->alu[1]->end,ref_trans->seq[alc->alu[1]->end],diff_trans->seq[alc->alu[1]->end],
neg,
pseudo
);
*/
ref_aa = ref_trans->seq[alc->alu[1]->end];
diff_aa = diff_trans->seq[alc->alu[1]->end];
if( ref_aa != diff_aa ) {
score_ratio += Probability2Score(tsm->unit[alc->alu[0]->end]->match_emission[ref_aa-'A']) - Probability2Score(tsm->unit[alc->alu[0]->end]->match_emission[diff_aa-'A']);
diff_score = Probability2Score(tsm->unit[alc->alu[0]->end]->match_emission[ref_aa-'A']) - Probability2Score(tsm->unit[alc->alu[0]->end]->match_emission[diff_aa-'A']);
if( diff_score < 0) {
count_ref_negative++;
if( diff_score > score_neg_5 ) {
count_ref_negative_0_5++;
} else if ( diff_score > score_neg_10 ) {
count_ref_negative_5_10++;
} else {
count_ref_negative_10_15++;
}
} else {
count_ref_positive++;
}
}
total_pseudo += pseudo;
total_neg += neg;
}
}
fprintf(ofp,"%s\t%s\t%.2f\t%d\t%d\t%d\t%d\t%d\n",ref->name,hmmp->baseseq->name,Score2Bits(score_ratio),
count_ref_positive,count_ref_negative,
count_ref_negative_0_5,
count_ref_negative_5_10,
count_ref_negative_10_15);
/*
fprintf(ofp,"%s,%s Total Pseudo %d vs Negative %d, Ratio %.4f Positive %d Negative %d Score %.2f Syn %d Changes %d NonSyn %d Changes %d\n",ref->name,hmmp->baseseq->name,total_pseudo,total_neg,Score2Bits(total_neg-total_pseudo),count_ref_positive,count_ref_negative,Score2Bits(score_ratio),syn_sites,syn_changes,nonsyn_sites,nonsyn_changes);
*/
free_Protein(hmmp);
}