PRIVATE double mfe_cost(const char *string, char *structure, const char *target)
{
#if TDIST
Tree *T1;
char *xstruc;
#endif
double energy, distance;
if (strlen(string)!=strlen(target)) {
fprintf(stderr, "%s\n%s\n", string, target);
nrerror("unequal length in mfe_cost");
}
energy = fold(string, structure);
#if TDIST
if (T0 == NULL) {
xstruc = expand_Full(target);
T0=make_tree(xstruc);
free(xstruc);
}
xstruc = expand_Full(structure);
T1=make_tree(xstruc);
distance = tree_edit_distance(T0,T1);
free(xstruc);
free_tree(T1);
#else
distance = (double) bp_distance(target, structure);
#endif
cost2 = energy_of_struct(string, target) - energy;
return (double) distance;
}
void print_path(char *seq, char *struc) {
int d;
char *s;
s = (char*)space((strlen(struc)+1)*sizeof(char));
strcpy(s, struc);
printf("%s\n%s %6.2f\n", seq, s, energy_of_struct(seq,s));
qsort(path, BP_dist, sizeof(move_t), compare_moves_when);
for (d=0; d<BP_dist; d++) {
int i,j;
i = path[d].i; j=path[d].j;
if (i<0) { /* delete */
s[(-i)-1] = s[(-j)-1] = '.';
} else {
s[i-1] = '('; s[j-1] = ')';
}
printf("%s %6.2f - %6.2f\n", s, energy_of_struct(seq,s), path[d].E/100.0);
}
free(s);
}
void ini_or_reset_rl(void) {
/* if there is no ringList-tree make a new one */
if (wurzl == NULL) {
ini_ringlist();
/* start structure */
struc2tree(GAV.startform);
GSV.currE = GSV.startE = energy_of_struct(GAV.farbe, GAV.startform);
/* stop structure(s) */
if ( GTV.stop ) {
int i;
qsort(GAV.stopform, GSV.maxS, sizeof(char *), comp_struc);
for (i = 0; i< GSV.maxS; i++)
GAV.sE[i] = energy_of_struct(GAV.farbe_full, GAV.stopform[i]);
}
else {
if(GTV.noLP)
noLonelyPairs=1;
initialize_fold(GSV.len);
/* fold sequence to get Minimum free energy structure (Mfe) */
GAV.sE[0] = fold(GAV.farbe_full, GAV.stopform[0]);
free_arrays();
/* revaluate energy of Mfe (maye differ if --logML=logarthmic */
GAV.sE[0] = energy_of_struct(GAV.farbe_full, GAV.stopform[0]);
}
GSV.stopE = GAV.sE[0];
ini_nbList(strlen(GAV.farbe_full)*strlen(GAV.farbe_full));
}
else {
/* reset ringlist-tree to start conditions */
reset_ringlist();
if(GTV.start) struc2tree(GAV.startform);
else {
GSV.currE = GSV.startE;
}
}
}
PRIVATE double pf_cost(const char *string, char *structure, const char *target)
{
#if PF
double f, e;
f = pf_fold(string, structure);
e = energy_of_struct(string, target);
return (double) (e-f-final_cost);
#else
nrerror("this version not linked with pf_fold");
return 0;
#endif
}
int main(int argc, char *argv[])
{
char *modelDir=NULL; /* Directory with model files */
struct svm_model* decision_model; /* SVM classification model */
/* Command line options */
int reverse=0; /* Scan reverse complement */
int showVersion=0; /* Shows version and exits */
int showHelp=0; /* Show short help and exits */
int from=-1; /* Scan slice from-to */
int to=-1;
FILE *clust_file=stdin; /* Input file */
FILE *out=stdout; /* Output file */
struct aln *AS[MAX_NUM_NAMES];
struct aln *window[MAX_NUM_NAMES];
char *tmpAln[MAX_NUM_NAMES];
int n_seq; /* number of input sequences */
int length; /* length of alignment/window */
int z_score_type;
int decision_model_type;
char *structure=NULL;
char *singleStruc,*gapStruc, *output,*woGapsSeq;
char strand[8];
char warningString[2000];
char warningString_regression[2000];
char *string=NULL;
double singleMFE,sumMFE,singleZ,sumZ,z,sci,id,decValue,prob,comb,entropy,GC;
double min_en, real_en;
int i,j,k,l,ll,r,countAln,nonGaps,singleGC;
int (*readFunction)(FILE *clust,struct aln *alignedSeqs[]);
char** lines=NULL;
int directions[3]={FORWARD,0,0};
int currDirection;
struct gengetopt_args_info args;
double meanMFE_fwd=0;
double consensusMFE_fwd=0;
double sci_fwd=0;
double z_fwd=0;
int strandGuess;
int avoid_shuffle=0;
double strandProb,strandDec;
if (cmdline_parser (argc, argv, &args) != 0){
usage();
exit(EXIT_FAILURE);
}
if (args.help_given){
help();
exit(EXIT_SUCCESS);
}
if (args.version_given){
version();
exit(EXIT_SUCCESS);
}
if (args.outfile_given){
out = fopen(args.outfile_arg, "w");
if (out == NULL){
fprintf(stderr, "ERROR: Can't open output file %s\n", args.outfile_arg);
exit(1);
}
}
/* Strand prediction implies both strands scored */
if (args.predict_strand_flag){
args.both_strands_flag=1;
}
if (args.forward_flag && !args.reverse_flag){
directions[0]=FORWARD;
directions[1]=directions[2]=0;
}
if (!args.forward_flag && args.reverse_flag){
directions[0]=REVERSE;
directions[1]=directions[2]=0;
}
if ((args.forward_flag && args.reverse_flag) || args.both_strands_flag){
directions[0]=FORWARD;
directions[1]=REVERSE;
}
if (args.window_given){
if (sscanf(args.window_arg,"%d-%d",&from,&to)!=2){
nrerror("ERROR: Invalid --window/-w command. "
"Use it like '--window 100-200'\n");
}
printf("from:%d,to:%d\n",from,to);
}
if (args.inputs_num>=1){
clust_file = fopen(args.inputs[0], "r");
if (clust_file == NULL){
fprintf(stderr, "ERROR: Can't open input file %s\n", args.inputs[0]);
exit(1);
}
}
/* Global RNA package variables */
do_backtrack = 1;
dangles=2;
switch(checkFormat(clust_file)){
case CLUSTAL:
readFunction=&read_clustal;
break;
case MAF:
readFunction=&read_maf;
break;
case 0:
nrerror("ERROR: Unknown alignment file format. Use Clustal W or MAF format.\n");
}
/* Set z-score type (mono/dinucleotide) here */
z_score_type = 2;
if (args.mononucleotide_given) z_score_type = 0;
/* now let's decide which decision model to take */
/* decision_model_type = 1 for normal model used in RNAz 1.0 */
/* decision_model_type = 2 for normal model using dinucelotide background */
/* decision_model_type = 3 for structural model using dinucelotide background */
decision_model_type = 2;
if (args.mononucleotide_given) decision_model_type = 1;
if (args.locarnate_given) decision_model_type = 3;
if ((args.mononucleotide_given) && args.locarnate_given){
z_score_type=2;
nrerror("ERROR: Structural decision model only trained with dinucleotide background model.\n");
}
if (args.no_shuffle_given) avoid_shuffle = 1;
decision_model=get_decision_model(NULL, decision_model_type);
/* Initialize Regression Models for mononucleotide */
/* Not needed if we score with dinucleotides */
if (z_score_type == 0) regression_svm_init();
countAln=0;
while ((n_seq=readFunction(clust_file, AS))!=0){
if (n_seq ==1){
nrerror("ERROR: You need at least two sequences in the alignment.\n");
}
countAln++;
length = (int) strlen(AS[0]->seq);
/* if a slice is specified by the user */
if ((from!=-1 || to!=-1) && (countAln==1)){
if ((from>=to)||(from<=0)||(to>length)){
nrerror("ERROR: Invalid window range given.\n");
}
sliceAln((const struct aln**)AS, (struct aln **)window, from, to);
length=to-from+1;
} else { /* take complete alignment */
/* window=AS does not work..., deep copy seems not necessary here*/
from=1;
to=length;
sliceAln((const struct aln **)AS, (struct aln **)window, 1, length);
}
/* Convert all Us to Ts for RNAalifold. There is a slight
difference in the results. During training we used alignments
with Ts, so we use Ts here as well. */
for (i=0;i<n_seq;i++){
j=0;
while (window[i]->seq[j]){
window[i]->seq[j]=toupper(window[i]->seq[j]);
if (window[i]->seq[j]=='U') window[i]->seq[j]='T';
++j;
}
}
k=0;
while ((currDirection=directions[k++])!=0){
if (currDirection==REVERSE){
revAln((struct aln **)window);
strcpy(strand,"reverse");
} else {
strcpy(strand,"forward");
}
structure = (char *) space((unsigned) length+1);
for (i=0;window[i]!=NULL;i++){
tmpAln[i]=window[i]->seq;
}
tmpAln[i]=NULL;
min_en = alifold(tmpAln, structure);
free_alifold_arrays();
comb=combPerPair(window,structure);
sumZ=0.0;
sumMFE=0.0;
GC=0.0;
output=(char *)space(sizeof(char)*(length+160)*(n_seq+1)*3);
strcpy(warningString,"");
strcpy(warningString_regression,"");
for (i=0;i<n_seq;i++){
singleStruc = space(strlen(window[i]->seq)+1);
woGapsSeq = space(strlen(window[i]->seq)+1);
j=0;
nonGaps=0;
singleGC=0;
while (window[i]->seq[j]){
/* Convert all Ts to Us for RNAfold. There is a difference
between the results. With U in the function call, we get
the results as RNAfold gives on the command line. Since
this variant was also used during training, we use it here
as well. */
if (window[i]->seq[j]=='T') window[i]->seq[j]='U';
if (window[i]->seq[j]=='C') singleGC++;
if (window[i]->seq[j]=='G') singleGC++;
if (window[i]->seq[j]!='-'){
nonGaps++;
woGapsSeq[strlen(woGapsSeq)]=window[i]->seq[j];
woGapsSeq[strlen(woGapsSeq)]='\0';
}
++j;
}
/* z-score is calculated here! */
singleMFE = fold(woGapsSeq, singleStruc);
free_arrays();
/* z-score type may be overwritten. If it is out of training
bounds, we switch to shuffling if allowed (avoid_shuffle). */
int z_score_type_orig = z_score_type;
singleZ=mfe_zscore(woGapsSeq,singleMFE, &z_score_type, avoid_shuffle, warningString_regression);
GC+=(double) singleGC/nonGaps;
sumZ+=singleZ;
sumMFE+=singleMFE;
if (window[1]->strand!='?' && !args.window_given){
sprintf(output+strlen(output),
">%s %d %d %c %d\n",
window[i]->name,window[i]->start,
window[i]->length,window[i]->strand,
window[i]->fullLength);
} else {
sprintf(output+strlen(output),">%s\n",window[i]->name);
}
gapStruc= (char *) space(sizeof(char)*(strlen(window[i]->seq)+1));
l=ll=0;
while (window[i]->seq[l]!='\0'){
if (window[i]->seq[l]!='-'){
gapStruc[l]=singleStruc[ll];
l++;
ll++;
} else {
gapStruc[l]='-';
l++;
}
}
char ch;
ch = 'R';
if (z_score_type == 1 || z_score_type == 3) ch = 'S';
sprintf(output+strlen(output),"%s\n%s ( %6.2f, z-score = %6.2f, %c)\n",
window[i]->seq,gapStruc,singleMFE,singleZ,ch);
z_score_type = z_score_type_orig;
free(woGapsSeq);
free(singleStruc);
}
{
int i; double s=0;
extern int eos_debug;
eos_debug=-1; /* shut off warnings about nonstandard pairs */
for (i=0; window[i]!=NULL; i++)
s += energy_of_struct(window[i]->seq, structure);
real_en = s/i;
}
string = consensus((const struct aln**) window);
sprintf(output+strlen(output),
">consensus\n%s\n%s (%6.2f = %6.2f + %6.2f) \n",
string, structure, min_en, real_en, min_en-real_en );
free(string);
id=meanPairID((const struct aln**)window);
entropy=NormShannonEntropy((const struct aln**)window);
z=sumZ/n_seq;
GC=(double)GC/n_seq;
if (sumMFE==0){
/*Set SCI to 0 in the weird case of no structure in single
sequences*/
sci=0;
} else {
sci=min_en/(sumMFE/n_seq);
}
decValue=999;
prob=0;
classify(&prob,&decValue,decision_model,id,n_seq,z,sci,entropy,decision_model_type);
if (args.cutoff_given){
if (prob<args.cutoff_arg){
continue;
}
}
warning(warningString,id,n_seq,z,sci,entropy,(struct aln **)window,decision_model_type);
fprintf(out,"\n############################ RNAz "PACKAGE_VERSION" ##############################\n\n");
fprintf(out," Sequences: %u\n", n_seq);
if (args.window_given){
fprintf(out," Slice: %u to %u\n",from,to);
}
fprintf(out," Columns: %u\n",length);
fprintf(out," Reading direction: %s\n",strand);
fprintf(out," Mean pairwise identity: %6.2f\n", id);
fprintf(out," Shannon entropy: %2.5f\n", entropy);
fprintf(out," G+C content: %2.5f\n", GC);
fprintf(out," Mean single sequence MFE: %6.2f\n", sumMFE/n_seq);
fprintf(out," Consensus MFE: %6.2f\n",min_en);
fprintf(out," Energy contribution: %6.2f\n",real_en);
fprintf(out," Covariance contribution: %6.2f\n",min_en-real_en);
fprintf(out," Combinations/Pair: %6.2f\n",comb);
fprintf(out," Mean z-score: %6.2f\n",z);
fprintf(out," Structure conservation index: %6.2f\n",sci);
if (decision_model_type == 1) {
fprintf(out," Background model: mononucleotide\n");
fprintf(out," Decision model: sequence based alignment quality\n");
}
if (decision_model_type == 2) {
fprintf(out," Background model: dinucleotide\n");
fprintf(out," Decision model: sequence based alignment quality\n");
}
if (decision_model_type == 3) {
fprintf(out," Background model: dinucleotide\n");
fprintf(out," Decision model: structural RNA alignment quality\n");
}
fprintf(out," SVM decision value: %6.2f\n",decValue);
fprintf(out," SVM RNA-class probability: %6f\n",prob);
if (prob>0.5){
fprintf(out," Prediction: RNA\n");
}
else {
fprintf(out," Prediction: OTHER\n");
}
fprintf(out,"%s",warningString_regression);
fprintf(out,"%s",warningString);
fprintf(out,"\n######################################################################\n\n");
fprintf(out,"%s",output);
fflush(out);
free(structure);
free(output);
if (currDirection==FORWARD && args.predict_strand_flag){
meanMFE_fwd=sumMFE/n_seq;
consensusMFE_fwd=min_en;
sci_fwd=sci;
z_fwd=z;
}
if (currDirection==REVERSE && args.predict_strand_flag){
if (predict_strand(sci_fwd-sci, meanMFE_fwd-(sumMFE/n_seq),
consensusMFE_fwd-min_en, z_fwd-z, n_seq, id,
&strandGuess, &strandProb, &strandDec, NULL)){
if (strandGuess==1){
fprintf(out, "\n# Strand winner: forward (%.2f)\n",strandProb);
} else {
fprintf(out, "\n# Strand winner: reverse (%.2f)\n",1-strandProb);
}
} else {
fprintf(out, "\n# WARNING: No strand prediction (values out of range)\n");
}
}
}
freeAln((struct aln **)AS);
freeAln((struct aln **)window);
}
if (args.inputs_num>=1){
fclose(clust_file);
}
cmdline_parser_free (&args);
if (countAln==0){
nrerror("ERROR: Empty alignment file\n");
}
svm_destroy_model(decision_model);
regression_svm_free();
return 0;
}
void print_stats(FILE* statsfile, char* seq, char* struc, int length, int iteration, int count_df_evaluations, double D, double prev_D, double norm, int printPS) {
plist *pl, *pl1,*pl2;
char fname[100];
char title[100];
char* ss;
double MEAgamma, mea, mea_en;
char* output;
int i,j;
static char timestamp[40];
const struct tm *tm;
time_t now;
ss = (char *) space((unsigned) length+1);
memset(ss,'.',length);
init_pf_fold(length);
pf_fold_pb(seq, NULL);
for (i = 1; i < length; i++) {
for (j = i+1; j<= length; j++) {
p_pp[i][j]=p_pp[j][i]=pr[iindx[i]-j];
}
}
get_pair_prob_vector(p_pp, p_unpaired, length, 1);
fprintf (stderr, "\nITERATION: %i\n", iteration);
fprintf(stderr, "DISCREPANCY: %.4f\n", D);
fprintf(stderr, "NORM: %.2f\n", norm);
if (prev_D > -1.0) {
fprintf(stderr, "IMPROVEMENT: %.4f%%\n\n", (1-(D/prev_D))*100);
}
fprintf(statsfile, "%i\t%.4f\t%.4f\t%i\t", iteration, D, norm, count_df_evaluations);
for (MEAgamma=1e-5; MEAgamma<1e+6; MEAgamma*=10 ) {
pl = make_plist(length, 1e-4/(1+MEAgamma));
mea = MEA(pl, ss, MEAgamma);
mea_en = energy_of_struct(seq, ss);
fprintf(statsfile,"%s,%.2e;", ss, MEAgamma);
free(pl);
}
fprintf(statsfile, "\t");
// Stochastic backtracking
fprintf(stderr, "Sampling structures...\n");
if (sample_structure) {
char* best_structure;
char* curr_structure;
double x;
double curr_energy = 0.0;
double min_energy = +1.0;
int curr_distance = 0;
int min_distance = 999999;
best_structure = (char *) space((unsigned) length+1);
for (i=1; i<=10000; i++) {
curr_structure = pbacktrack_pb(seq);
curr_energy = energy_of_struct(seq, curr_structure);
curr_distance = 0.0;
//fprintf(stderr, "%s%.2f ", curr_structure, curr_energy);
for (j = 1; j <= length; j++) {
if (q_unpaired[j] > -0.5) {
x = (curr_structure[j-1] == '.') ? 1.0 : 0.0;
curr_distance += abs(x-q_unpaired[j]);
}
}
if (curr_distance < min_distance) {
min_distance = curr_distance;
min_energy = curr_energy;
strcpy(best_structure, curr_structure);
}
if (curr_distance == min_distance) {
if (curr_energy < min_energy) {
min_energy = curr_energy;
strcpy(best_structure, curr_structure);
}
}
//fprintf(stderr, "%i\n", curr_distance);
free(curr_structure);
}
//fprintf(stderr, "\n%s %.2f %i\n", best_structure, min_energy, min_distance);
fprintf(statsfile, "\t%s\t%.2f\t%i\t", best_structure, min_energy, min_distance);
} else {
fprintf(statsfile, "NA\tNA\tNA\t");
}
for (i = 1; i <= length; i++) {
fprintf(statsfile, "%.4f", epsilon[i]);
if (!(i==length)) {
fprintf(statsfile, ",");
}
}
now = time ( NULL );
tm = localtime ( &now );
strftime ( timestamp, 40, "%Y-%m-%d %X", tm );
fprintf(statsfile, "\t%s\n", timestamp);
/* Print dotplot only if not noPS is given and function call asks for it */
if (!noPS && printPS) {
/* Print dotplot */
sprintf(fname,"%s/iteration%i.ps", psDir, iteration);
pl1 = make_plist(length, 1e-5);
if (struc != NULL) {
pl2 = b2plist(struc);
} else {
pl2 = NULL;
}
sprintf(title,"Iteration %i, D = %.4f", iteration, D);
(void) PS_dot_plot_list_epsilon(seq, fname, pl2, pl1, epsilon, title);
}
free_pf_arrays();
}
int main(int argc, char *argv[])
{
char *string/*, *line*/;
char *structure=NULL, *cstruc=NULL;
/*char fname[13], ffname[20], gfname[20];*/
/*char *ParamFile=NULL;*/
char *ns_bases=NULL, *c;
int i, length, l, sym/*, r*/;
double energy, min_en;
double kT, sfact=1.07;
int pf=0, noPS=0, istty;
int noconv=0;
int circ=0;
AjPSeq seq = NULL;
AjPFile confile = NULL;
AjPFile paramfile = NULL;
AjPFile outf = NULL;
AjPFile essfile = NULL;
AjPFile dotfilea = NULL;
AjPFile dotfileb = NULL;
AjPStr seqstring = NULL;
AjPStr constring = NULL;
AjPStr seqname = NULL;
float eT = 0.;
AjBool eGU;
AjBool ecirc = ajFalse;
AjBool eclose;
AjBool lonely;
AjBool convert;
AjPStr ensbases = NULL;
AjBool etloop;
AjPStr eenergy = NULL;
char ewt = '\0';
float escale = 0.;
AjPStr edangles = NULL;
char edangle = '\0';
ajint len;
embInitPV("vrnafold",argc,argv,"VIENNA",VERSION);
seqstring = ajStrNew();
constring = ajStrNew();
seqname = ajStrNew();
seq = ajAcdGetSeq("sequence");
confile = ajAcdGetInfile("constraintfile");
paramfile = ajAcdGetInfile("paramfile");
eT = ajAcdGetFloat("temperature");
ecirc = ajAcdGetBoolean("circular");
eGU = ajAcdGetBoolean("gu");
eclose = ajAcdGetBoolean("closegu");
lonely = ajAcdGetBoolean("lp");
convert = ajAcdGetBoolean("convert");
ensbases = ajAcdGetString("nsbases");
etloop = ajAcdGetBoolean("tetraloop");
eenergy = ajAcdGetListSingle("energy");
escale = ajAcdGetFloat("scale");
edangles = ajAcdGetListSingle("dangles");
outf = ajAcdGetOutfile("outfile");
essfile = ajAcdGetOutfile("ssoutfile");
/*
dotfilea = ajAcdGetOutfile("adotoutfile");
dotfileb = ajAcdGetOutfile("bdotoutfile");
*/
do_backtrack = 2;
pf = 0;
string = NULL;
istty = 0;
temperature = (double) eT;
circ = !!ecirc;
noGU = (eGU) ? 0 : 1;
no_closingGU = (eclose) ? 0 : 1;
noLonelyPairs = (lonely) ? 0 : 1;
noconv = (convert) ? 0 : 1;
ns_bases = (ajStrGetLen(ensbases)) ? MAJSTRGETPTR(ensbases) : NULL;
tetra_loop = !!etloop;
ewt = *ajStrGetPtr(eenergy);
if(ewt == '0')
energy_set = 0;
else if(ewt == '1')
energy_set = 1;
else if(ewt == '2')
energy_set = 2;
sfact = (double) escale;
edangle = *ajStrGetPtr(edangles);
if(edangle == '0')
dangles = 0;
else if(edangle == '1')
dangles = 1;
else if(edangle == '2')
dangles = 2;
else if(edangle == '3')
dangles = 3;
if(circ && noLonelyPairs)
{
ajWarn("Depending on the origin of the circular sequence\n"
"some structures may be missed when using -noLP\nTry "
"rotating your sequence a few times\n");
}
if(paramfile)
read_parameter_file(paramfile);
if (ns_bases != NULL)
{
nonstandards = space(33);
c=ns_bases;
i=sym=0;
if (*c=='-')
{
sym=1; c++;
}
while (*c!='\0')
{
if (*c!=',')
{
nonstandards[i++]=*c++;
nonstandards[i++]=*c;
if ((sym)&&(*c!=*(c-1)))
{
nonstandards[i++]=*c;
nonstandards[i++]=*(c-1);
}
}
c++;
}
}
if(confile)
vienna_GetConstraints(confile,&constring);
string = NULL;
structure = NULL;
length = ajSeqGetLen(seq);
string = (char *) space(length+1);
strcpy(string,ajSeqGetSeqC(seq));
len = ajStrGetLen(constring);
structure = (char *) space(length+1);
if(len)
{
fold_constrained = 1;
strcpy(structure,ajStrGetPtr(constring));
}
for (l = 0; l < length; l++) {
string[l] = toupper(string[l]);
if (!noconv && string[l] == 'T') string[l] = 'U';
}
/* initialize_fold(length); */
if (circ)
min_en = circfold(string, structure);
else
min_en = fold(string, structure);
ajFmtPrintF(outf,"%s\n%s", string, structure);
if (istty)
printf("\n minimum free energy = %6.2f kcal/mol\n", min_en);
else
ajFmtPrintF(outf," (%6.2f)\n", min_en);
if (!noPS)
{
if (length<2000)
(void) PS_rna_plot(string, structure, essfile);
else
ajWarn("Structure too long, not doing xy_plot\n");
}
if (length>=2000) free_arrays();
if (pf)
{
char *pf_struc;
pf_struc = (char *) space((unsigned) length+1);
if (dangles==1)
{
dangles=2; /* recompute with dangles as in pf_fold() */
min_en = (circ) ? energy_of_circ_struct(string, structure) :
energy_of_struct(string, structure);
dangles=1;
}
kT = (temperature+273.15)*1.98717/1000.; /* in Kcal */
pf_scale = exp(-(sfact*min_en)/kT/length);
if (length>2000)
ajWarn("scaling factor %f\n", pf_scale);
(circ) ? init_pf_circ_fold(length) : init_pf_fold(length);
if (cstruc!=NULL)
strncpy(pf_struc, cstruc, length+1);
energy = (circ) ? pf_circ_fold(string, pf_struc) :
pf_fold(string, pf_struc);
if (do_backtrack)
{
ajFmtPrintF(outf,"%s", pf_struc);
ajFmtPrintF(outf," [%6.2f]\n", energy);
}
if ((istty)||(!do_backtrack))
ajFmtPrintF(outf," free energy of ensemble = %6.2f kcal/mol\n",
energy);
if (do_backtrack)
{
plist *pl1,*pl2;
char *cent;
double dist, cent_en;
cent = centroid(length, &dist);
cent_en = (circ) ? energy_of_circ_struct(string, cent) :
energy_of_struct(string, cent);
ajFmtPrintF(outf,"%s {%6.2f d=%.2f}\n", cent, cent_en, dist);
free(cent);
pl1 = make_plist(length, 1e-5);
pl2 = b2plist(structure);
(void) PS_dot_plot_list(string, dotfilea, pl1, pl2, "");
free(pl2);
if (do_backtrack==2)
{
pl2 = stackProb(1e-5);
PS_dot_plot_list(string, dotfileb, pl1, pl2,
"Probabilities for stacked pairs (i,j)(i+1,j-1)");
free(pl2);
}
free(pl1);
free(pf_struc);
}
ajFmtPrintF(outf," frequency of mfe structure in ensemble %g; ",
exp((energy-min_en)/kT));
if (do_backtrack)
ajFmtPrintF(outf,"ensemble diversity %-6.2f", mean_bp_dist(length));
ajFmtPrintF(outf,"\n");
free_pf_arrays();
}
if (cstruc!=NULL)
free(cstruc);
free(string);
free(structure);
ajStrDel(&seqstring);
ajStrDel(&constring);
ajStrDel(&seqname);
ajStrDel(&ensbases);
ajStrDel(&eenergy);
ajStrDel(&edangles);
ajSeqDel(&seq);
ajFileClose(&confile);
ajFileClose(¶mfile);
ajFileClose(&outf);
ajFileClose(&essfile);
/*
ajFileClose(&dotfilea);
ajFileClose(&dotfileb);
*/
if (length<2000) free_arrays();
embExit();
return 0;
}
int main(int argc, char *argv[])
{
char *string/*, *line*/;
char *structure=NULL, *cstruc=NULL;
/*char fname[53], ffname[60]; */
/*char *ParamFile=NULL; */
char *ns_bases=NULL, *c;
char *Concfile;
int i, length, l, sym/*, r*/;
double min_en;
double kT, sfact=1.07;
int pf=0, istty;
int noconv=0;
int doT=0; /*compute dimere free energies etc.*/
int doC=0; /*toggle to compute concentrations*/
int doQ=0; /*toggle to compute prob of base being paired*/
int cofi=0; /*toggle concentrations stdin / file*/
struct plist *prAB;
struct plist *prAA; /*pair probabilities of AA dimer*/
struct plist *prBB;
struct plist *prA;
struct plist *prB;
struct plist *mfAB;
struct plist *mfAA; /*pair mfobabilities of AA dimer*/
struct plist *mfBB;
struct plist *mfA;
struct plist *mfB;
double *ConcAandB;
AjPSeq seq1 = NULL;
AjPFile confile1 = NULL;
AjPSeq seq2 = NULL;
AjPFile confile2 = NULL;
AjPFile concfile = NULL;
AjPFile paramfile = NULL;
AjPFile outf = NULL;
AjPFile essfile = NULL;
AjPFile dotfile = NULL;
AjPFile aoutf = NULL;
AjPFile aaoutf = NULL;
AjPFile boutf = NULL;
AjPFile bboutf = NULL;
AjPFile aboutf = NULL;
AjPStr seqstring1 = NULL;
AjPStr constring1 = NULL;
AjPStr constring2 = NULL;
float eT = 0.;
AjBool eGU;
AjBool eclose;
AjBool lonely;
AjBool convert;
AjPStr ensbases = NULL;
AjBool etloop;
AjPStr eenergy = NULL;
char ewt = '\0';
float escale = 0.;
AjPStr edangles = NULL;
char edangle = '\0';
/* AjBool dimers; */
/* AjBool paired; */
embInitPV("vrnacofold",argc,argv,"VIENNA",VERSION);
seqstring1 = ajStrNew();
constring1 = ajStrNew();
constring2 = ajStrNew();
seq1 = ajAcdGetSeq("asequence");
confile1 = ajAcdGetInfile("aconstraintfile");
seq2 = ajAcdGetSeq("bsequence");
confile2 = ajAcdGetInfile("bconstraintfile");
paramfile = ajAcdGetInfile("paramfile");
eT = ajAcdGetFloat("temperature");
eGU = ajAcdGetBoolean("gu");
eclose = ajAcdGetBoolean("closegu");
lonely = ajAcdGetBoolean("lp");
convert = ajAcdGetBoolean("convert");
ensbases = ajAcdGetString("nsbases");
etloop = ajAcdGetBoolean("tetraloop");
eenergy = ajAcdGetListSingle("energy");
escale = ajAcdGetFloat("scale");
edangles = ajAcdGetListSingle("dangles");
/* dimers = ajAcdGetBoolean("dimers"); */
/* paired = ajAcdGetBoolean("paired"); */
outf = ajAcdGetOutfile("outfile");
essfile = ajAcdGetOutfile("ssoutfile");
/* concfile = ajAcdGetInfile("concentrationfile"); */
/* dotfile = ajAcdGetOutfile("dotoutfile"); */
/*
aoutf = ajAcdGetOutfile("aoutfile");
aaoutf = ajAcdGetOutfile("aaoutfile");
boutf = ajAcdGetOutfile("boutfile");
bboutf = ajAcdGetOutfile("bboutfile");
aboutf = ajAcdGetOutfile("aboutfile");
*/
do_backtrack = 1;
pf = 0;
doT = 0;
doC = 0;
cofi = 0;
doQ = 0;
string = NULL;
Concfile = NULL;
istty = 0;
temperature = (double) eT;
noGU = (eGU) ? 0 : 1;
no_closingGU = (eclose) ? 0 : 1;
noLonelyPairs = (lonely) ? 0 : 1;
noconv = (convert) ? 0 : 1;
ns_bases = (ajStrGetLen(ensbases)) ? MAJSTRGETPTR(ensbases) : NULL;
tetra_loop = !!etloop;
ewt = *ajStrGetPtr(eenergy);
if(ewt == '0')
energy_set = 0;
else if(ewt == '1')
energy_set = 1;
else if(ewt == '2')
energy_set = 2;
sfact = (double) escale;
edangle = *ajStrGetPtr(edangles);
if(edangle == '0')
dangles = 0;
else if(edangle == '1')
dangles = 1;
else if(edangle == '2')
dangles = 2;
else if(edangle == '3')
dangles = 3;
if(paramfile)
read_parameter_file(paramfile);
if (ns_bases != NULL)
{
nonstandards = space(33);
c=ns_bases;
i=sym=0;
if (*c=='-')
{
sym=1; c++;
}
while (*c!='\0')
{
if (*c!=',')
{
nonstandards[i++]=*c++;
nonstandards[i++]=*c;
if ((sym)&&(*c!=*(c-1)))
{
nonstandards[i++]=*c;
nonstandards[i++]=*(c-1);
}
}
c++;
}
}
cut_point = -1;
ajFmtPrintS(&seqstring1,"%s&%s",ajSeqGetSeqC(seq1),ajSeqGetSeqC(seq2));
string = tokenize(MAJSTRGETPTR(seqstring1)); /* frees line */
length = (int) strlen(string);
if (doC)
{
ConcAandB = read_concentrations(concfile);
}
structure = (char *) space((unsigned) length+1);
if(confile1)
{
vienna_GetConstraints(confile1,&constring1);
vienna_GetConstraints(confile2,&constring2);
ajStrAppendK(&constring1,'&');
ajStrAppendS(&constring1,constring2);
cstruc = tokenize(MAJSTRGETPTR(constring1));
if (cstruc!=NULL)
strncpy(structure, cstruc, length);
else
ajFatal("Constraints missing\n");
}
for (l = 0; l < length; l++)
{
string[l] = toupper(string[l]);
if (!noconv && string[l] == 'T') string[l] = 'U';
}
/*compute mfe of AB dimer*/
min_en = cofold(string, structure);
mfAB=(struct plist *) space(sizeof(struct plist) * (length+1));
mfAB=get_mfe_plist(mfAB);
if (cut_point == -1)
ajFmtPrintF(outf,"%s\n%s", string, structure); /*no cofold*/
else
{
char *pstring, *pstruct;
pstring = costring(string);
pstruct = costring(structure);
ajFmtPrintF(outf,"%s\n%s", pstring, pstruct);
free(pstring);
free(pstruct);
}
ajFmtPrintF(outf," (%6.2f)\n", min_en);
if (length<2000)
(void) PS_rna_plot(string, structure, essfile);
else
{
ajWarn("Structure too long, not doing xy_plot\n");
free_co_arrays();
}
/*compute partition function*/
if (pf)
{
cofoldF AB, AA, BB;
if (dangles==1)
{
dangles=2; /* recompute with dangles as in pf_fold() */
min_en = energy_of_struct(string, structure);
dangles=1;
}
kT = (temperature+273.15)*1.98717/1000.; /* in Kcal */
pf_scale = exp(-(sfact*min_en)/kT/length);
if (length>2000)
ajWarn("scaling factor %f\n", pf_scale);
init_co_pf_fold(length);
if (cstruc!=NULL)
strncpy(structure, cstruc, length+1);
AB = co_pf_fold(string, structure);
if (do_backtrack)
{
char *costruc;
costruc = (char *) space(sizeof(char)*(strlen(structure)+2));
if (cut_point<0)
ajFmtPrintF(outf,"%s", structure);
else
{
strncpy(costruc, structure, cut_point-1);
strcat(costruc, "&");
strcat(costruc, structure+cut_point-1);
ajFmtPrintF(outf,"%s", costruc);
}
ajFmtPrintF(outf," [%6.2f]\n", AB.FAB);
}
if ((istty)||(!do_backtrack))
ajFmtPrintF(outf," free energy of ensemble = %6.2f kcal/mol\n",
AB.FAB);
ajFmtPrintF(outf," frequency of mfe structure in ensemble %g",
exp((AB.FAB-min_en)/kT));
ajFmtPrintF(outf," , delta G binding=%6.2f\n", AB.FcAB - AB.FA - AB.FB);
prAB=(struct plist *) space(sizeof(struct plist) * (2*length));
prAB=get_plist(prAB, length,0.00001);
/* if (doQ) make_probsum(length,fname); */ /*compute prob of base paired*/
/* free_co_arrays(); */
if (doT)
{ /* cofold of all dimers, monomers */
int Blength, Alength;
char *Astring, *Bstring;
char *Newstring;
/*char Newname[30];*/
char comment[80];
if (cut_point<0)
{
free(mfAB);
free(prAB);
ajFatal("Sorry, I cannot do that with only one molecule, "
"please give me two\n");
}
if (dangles==1)
dangles=2;
Alength=cut_point-1; /*length of first molecule*/
Blength=length-cut_point+1; /*length of 2nd molecule*/
/*Sequence of first molecule*/
Astring=(char *)space(sizeof(char)*(Alength+1));
/*Sequence of second molecule*/
Bstring=(char *)space(sizeof(char)*(Blength+1));
strncat(Astring,string,Alength);
strncat(Bstring,string+Alength,Blength);
/* compute AA dimer */
prAA=(struct plist *) space(sizeof(struct plist) * (4*Alength));
mfAA=(struct plist *) space(sizeof(struct plist) * (Alength+1));
AA=do_partfunc(Astring, Alength, 2, &prAA, &mfAA);
/* compute BB dimer */
prBB=(struct plist *) space(sizeof(struct plist) * (4*Blength));
mfBB=(struct plist *) space(sizeof(struct plist) * (Blength+1));
BB=do_partfunc(Bstring, Blength, 2, &prBB, &mfBB);
/*free_co_pf_arrays();*/
/* compute A monomer */
prA=(struct plist *) space(sizeof(struct plist) * (2*Alength));
mfA=(struct plist *) space(sizeof(struct plist) * (Alength+1));
do_partfunc(Astring, Alength, 1, &prA, &mfA);
/* compute B monomer */
prB=(struct plist *) space(sizeof(struct plist) * (2*Blength));
mfB=(struct plist *) space(sizeof(struct plist) * (Blength+1));
do_partfunc(Bstring, Blength, 1, &prB, &mfB);
compute_probabilities(AB.F0AB, AB.FA, AB.FB, prAB, prA, prB,
Alength);
compute_probabilities(AA.F0AB, AA.FA, AA.FA, prAA, prA, prA,
Alength);
compute_probabilities(BB.F0AB, BB.FA, BB.FA, prBB, prA, prB,
Blength);
ajFmtPrintF(outf,"Free Energies:\nAB\t\tAA\t\tBB\t\tA\t\tB\n%.6f"
"\t%6f\t%6f\t%6f\t%6f\n",
AB.FcAB, AA.FcAB, BB.FcAB, AB.FA, AB.FB);
if (doC)
{
do_concentrations(AB.FcAB, AA.FcAB, BB.FcAB, AB.FA, AB.FB,
ConcAandB, outf);
free(ConcAandB);/*freeen*/
}
/*AB dot_plot*/
/*write Free Energy into comment*/
sprintf(comment,"\n%%Heterodimer AB FreeEnergy= %.9f\n", AB.FcAB);
/*reset cut_point*/
cut_point=Alength+1;
(void)PS_dot_plot_list(string, aboutf, prAB, mfAB, comment);
/*AA dot_plot*/
sprintf(comment,"\n%%Homodimer AA FreeEnergy= %.9f\n",AA.FcAB);
/*write AA sequence*/
Newstring=(char*)space((2*Alength+1)*sizeof(char));
strcpy(Newstring,Astring);
strcat(Newstring,Astring);
(void)PS_dot_plot_list(Newstring, aaoutf, prAA, mfAA, comment);
free(Newstring);
/*BB dot_plot*/
sprintf(comment,"\n%%Homodimer BB FreeEnergy= %.9f\n",BB.FcAB);
/*write BB sequence*/
Newstring=(char*)space((2*Blength+1)*sizeof(char));
strcpy(Newstring,Bstring);
strcat(Newstring,Bstring);
/*reset cut_point*/
cut_point=Blength+1;
(void)PS_dot_plot_list(Newstring, bboutf, prBB, mfBB, comment);
free(Newstring);
/*A dot plot*/
/*reset cut_point*/
cut_point=-1;
sprintf(comment,"\n%%Monomer A FreeEnergy= %.9f\n",AB.FA);
/*write A sequence*/
(void)PS_dot_plot_list(Astring, aoutf, prA, mfA, comment);
/*B monomer dot plot*/
sprintf(comment,"\n%%Monomer B FreeEnergy= %.9f\n",AB.FB);
/*write B sequence*/
(void)PS_dot_plot_list(Bstring, boutf, prB, mfB, comment);
free(Astring);
free(Bstring);
free(prAB);
free(prAA);
free(prBB);
free(prA);
free(prB);
free(mfAB);
free(mfAA);
free(mfBB);
free(mfA);
free(mfB);
} /*end if(doT)*/
}/*end if(pf)*/
if (do_backtrack)
{
if (!doT)
{
if (pf)
{
(void) PS_dot_plot_list(string, dotfile, prAB, mfAB, "doof");
free(prAB);
}
free(mfAB);
}
}
if (!doT)
free_co_pf_arrays();
if (cstruc!=NULL)
free(cstruc);
free(string);
free(structure);
ajStrDel(&seqstring1);
ajStrDel(&constring1);
ajStrDel(&constring2);
ajSeqDel(&seq1);
ajSeqDel(&seq2);
ajStrDel(&ensbases);
ajStrDel(&eenergy);
ajStrDel(&edangles);
ajFileClose(&confile1);
ajFileClose(&confile2);
ajFileClose(¶mfile);
ajFileClose(&outf);
ajFileClose(&essfile);
if (length<2000)
free_co_arrays();
embExit();
return 0;
}
path_t* get_path(char *seq, char *s1, char* s2,
int maxkeep, int *num_entry) {
int E, d;
path_t *route;
E = find_saddle(seq, s1, s2, maxkeep);
*num_entry = BP_dist+1;
route = (path_t *)space((BP_dist+1)*sizeof(path_t));
qsort(path, BP_dist, sizeof(move_t), compare_moves_when);
if (path_fwd) {
/* memorize start of path */
route[0].s = (char*)space((strlen(s1)+1)*sizeof(char));
strcpy(route[0].s, s1);
route[0].en = energy_of_struct(seq, s1);
for (d=0; d<BP_dist; d++) {
int i,j;
route[d+1].s = (char*)space((strlen(route[d].s)+1)*sizeof(char));
strcpy(route[d+1].s, route[d].s);
i = path[d].i; j=path[d].j;
if (i<0) { /* delete */
route[d+1].s[(-i)-1] = route[d+1].s[(-j)-1] = '.';
} else {
route[d+1].s[i-1] = '('; route[d+1].s[j-1] = ')';
}
route[d+1].en = path[d].E/100.0;
}
}
else {
/* memorize start of path */
route[BP_dist].s = (char*)space((strlen(s2)+1)*sizeof(char));
strcpy(route[BP_dist].s, s2);
route[BP_dist].en = energy_of_struct(seq, s2);
for (d=0; d<BP_dist; d++) {
int i,j;
route[BP_dist-d-1].s
= (char*)space((strlen(route[BP_dist-d].s)+1)*sizeof(char));
strcpy(route[BP_dist-d-1].s, route[BP_dist-d].s);
i = path[d].i; j=path[d].j;
if (i<0) { /* delete */
route[BP_dist-d-1].s[(-i)-1] = route[BP_dist-d-1].s[(-j)-1] = '.';
} else {
route[BP_dist-d-1].s[i-1] = '('; route[BP_dist-d-1].s[j-1] = ')';
}
route[BP_dist-d-1].en = path[d].E/100.0;
}
}
#if _DEBUG_FINDPATH_
fprintf(stderr, "\n%s\n%s\n%s\n\n", seq, s1, s2);
for (d=0; d<=BP_dist; d++)
fprintf(stderr, "%s %6.2f\n", route[d].s, route[d].en);
fprintf(stderr, "%d\n", *num_entry);
#endif
free(path);
path_fwd = 0;
return (route);
}
int main(int argc, char *argv[])
{
char *string, *line;
char *structure=NULL, *cstruc=NULL;
char fname[13], ffname[20], gfname[20];
char *ParamFile=NULL;
char *ns_bases=NULL, *c;
int i, length, l, sym, r;
double energy, min_en;
double kT, sfact=1.07;
int pf=0, noPS=0, istty;
int noconv=0;
int circ=0;
do_backtrack = 1;
string=NULL;
for (i=1; i<argc; i++) {
if (argv[i][0]=='-')
switch ( argv[i][1] )
{
case 'T': if (argv[i][2]!='\0') usage();
if(i==argc-1) usage();
r=sscanf(argv[++i], "%lf", &temperature);
if (!r) usage();
break;
case 'p': pf=1;
if (argv[i][2]!='\0')
(void) sscanf(argv[i]+2, "%d", &do_backtrack);
break;
case 'n':
if ( strcmp(argv[i], "-noGU")==0) noGU=1;
if ( strcmp(argv[i], "-noCloseGU")==0) no_closingGU=1;
if ( strcmp(argv[i], "-noLP")==0) noLonelyPairs=1;
if ( strcmp(argv[i], "-noPS")==0) noPS=1;
if ( strcmp(argv[i], "-nsp") ==0) {
if (i==argc-1) usage();
ns_bases = argv[++i];
}
if ( strcmp(argv[i], "-noconv")==0) noconv=1;
break;
case '4':
tetra_loop=0;
break;
case 'e':
if(i==argc-1) usage();
r=sscanf(argv[++i],"%d", &energy_set);
if (!r) usage();
break;
case 'C':
fold_constrained=1;
break;
case 'c':
if ( strcmp(argv[i], "-circ")==0) circ=1;
break;
case 'S':
if(i==argc-1) usage();
r=sscanf(argv[++i],"%lf", &sfact);
if (!r) usage();
break;
case 'd': dangles=0;
if (argv[i][2]!='\0') {
r=sscanf(argv[i]+2, "%d", &dangles);
if (r!=1) usage();
}
break;
case 'P':
if (i==argc-1) usage();
ParamFile = argv[++i];
break;
default: usage();
}
}
if (circ && noLonelyPairs)
fprintf(stderr, "warning, depending on the origin of the circular sequence, some structures may be missed when using -noLP\nTry rotating your sequence a few times\n");
if (ParamFile != NULL)
read_parameter_file(ParamFile);
if (ns_bases != NULL) {
nonstandards = space(33);
c=ns_bases;
i=sym=0;
if (*c=='-') {
sym=1; c++;
}
while (*c!='\0') {
if (*c!=',') {
nonstandards[i++]=*c++;
nonstandards[i++]=*c;
if ((sym)&&(*c!=*(c-1))) {
nonstandards[i++]=*c;
nonstandards[i++]=*(c-1);
}
}
c++;
}
}
istty = isatty(fileno(stdout))&&isatty(fileno(stdin));
if ((fold_constrained)&&(istty)) {
printf("Input constraints using the following notation:\n");
printf("| : paired with another base\n");
printf(". : no constraint at all\n");
printf("x : base must not pair\n");
printf("< : base i is paired with a base j<i\n");
printf("> : base i is paired with a base j>i\n");
printf("matching brackets ( ): base i pairs base j\n");
}
do { /* main loop: continue until end of file */
if (istty) {
printf("\nInput string (upper or lower case); @ to quit\n");
printf("%s%s\n", scale1, scale2);
}
fname[0]='\0';
if ((line = get_line(stdin))==NULL) break;
/* skip comment lines and get filenames */
while ((*line=='*')||(*line=='\0')||(*line=='>')) {
if (*line=='>')
(void) sscanf(line, ">%12s", fname);
printf("%s\n", line);
free(line);
if ((line = get_line(stdin))==NULL) break;
}
if ((line ==NULL) || (strcmp(line, "@") == 0)) break;
string = (char *) space(strlen(line)+1);
(void) sscanf(line,"%s",string);
free(line);
length = (int) strlen(string);
structure = (char *) space((unsigned) length+1);
if (fold_constrained) {
cstruc = get_line(stdin);
if (cstruc!=NULL)
strncpy(structure, cstruc, length);
else
fprintf(stderr, "constraints missing\n");
}
for (l = 0; l < length; l++) {
string[l] = toupper(string[l]);
if (!noconv && string[l] == 'T') string[l] = 'U';
}
if (istty)
printf("length = %d\n", length);
/* initialize_fold(length); */
if (circ)
min_en = circfold(string, structure);
else
min_en = fold(string, structure);
printf("%s\n%s", string, structure);
if (istty)
printf("\n minimum free energy = %6.2f kcal/mol\n", min_en);
else
printf(" (%6.2f)\n", min_en);
(void) fflush(stdout);
if (fname[0]!='\0') {
strcpy(ffname, fname);
strcat(ffname, "_ss.ps");
strcpy(gfname, fname);
strcat(gfname, "_ss.g");
} else {
strcpy(ffname, "rna.ps");
strcpy(gfname, "rna.g");
}
if (!noPS) {
if (length<2000)
(void) PS_rna_plot(string, structure, ffname);
else
fprintf(stderr,"INFO: structure too long, not doing xy_plot\n");
}
if (length>2000) free_arrays();
if (pf) {
char *pf_struc;
pf_struc = (char *) space((unsigned) length+1);
if (dangles==1) {
dangles=2; /* recompute with dangles as in pf_fold() */
min_en = (circ) ? energy_of_circ_struct(string, structure) : energy_of_struct(string, structure);
dangles=1;
}
kT = (temperature+273.15)*1.98717/1000.; /* in Kcal */
pf_scale = exp(-(sfact*min_en)/kT/length);
if (length>2000) fprintf(stderr, "scaling factor %f\n", pf_scale);
(circ) ? init_pf_circ_fold(length) : init_pf_fold(length);
if (cstruc!=NULL)
strncpy(pf_struc, cstruc, length+1);
energy = (circ) ? pf_circ_fold(string, pf_struc) : pf_fold(string, pf_struc);
if (do_backtrack) {
printf("%s", pf_struc);
if (!istty) printf(" [%6.2f]\n", energy);
else printf("\n");
}
if ((istty)||(!do_backtrack))
printf(" free energy of ensemble = %6.2f kcal/mol\n", energy);
if (do_backtrack) {
plist *pl1,*pl2;
char *cent;
double dist, cent_en;
cent = centroid(length, &dist);
cent_en = (circ) ? energy_of_circ_struct(string, cent) :energy_of_struct(string, cent);
printf("%s {%6.2f d=%.2f}\n", cent, cent_en, dist);
free(cent);
if (fname[0]!='\0') {
strcpy(ffname, fname);
strcat(ffname, "_dp.ps");
} else strcpy(ffname, "dot.ps");
pl1 = make_plist(length, 1e-5);
pl2 = b2plist(structure);
(void) PS_dot_plot_list(string, ffname, pl1, pl2, "");
free(pl2);
if (do_backtrack==2) {
pl2 = stackProb(1e-5);
if (fname[0]!='\0') {
strcpy(ffname, fname);
strcat(ffname, "_dp2.ps");
} else strcpy(ffname, "dot2.ps");
PS_dot_plot_list(string, ffname, pl1, pl2,
"Probabilities for stacked pairs (i,j)(i+1,j-1)");
free(pl2);
}
free(pl1);
free(pf_struc);
}
printf(" frequency of mfe structure in ensemble %g; ",
exp((energy-min_en)/kT));
if (do_backtrack)
printf("ensemble diversity %-6.2f", mean_bp_dist(length));
printf("\n");
free_pf_arrays();
}
if (cstruc!=NULL) free(cstruc);
(void) fflush(stdout);
free(string);
free(structure);
} while (1);
return 0;
}
/* PURPOSE: Sets the predictor variables
PARAMS: char *window: Window of alignment for which strand is predicted
char *window_reverse: Reverse complement of window
int length: Length of alignment window
int n_seq: Number of sequences in alignment window
double *deltaSCI: 1st predictor
double *deltaMeanMFE: 2nd predictor
double *deltaConsMFE: 3rd predictor
double *deltaZ: 4th predictor
char *modelDir: Directory with SVM models
RETURNS: 0 on success
*/
void get_strand_predictors(
char *window[],
char *window_reverse[],
int length,
int n_seq,
double *deltaSCI,
double *deltaMeanMFE,
double *deltaConsMFE,
double *deltaZ,
char *modelDir)
{
int i,j;
char *structure=NULL, *structure_reverse=NULL;
char *singleStruc, *woGapsSeq, *singleStruc_reverse, *woGapsSeq_reverse;
double decValue=0;
double singleMFE,sumMFE,singleZ,sumZ,z,sci;
double singleMFE_reverse,sumMFE_reverse,singleZ_reverse,sumZ_reverse,z_reverse,sci_reverse;
double min_en, real_en;
double min_en_reverse, real_en_reverse;
int z_score_type;
char warningString_strand[2000];
sumZ=0; sumMFE=0, sumZ_reverse=0, sumMFE_reverse=0;
/* RNAstrand is only trained on mononucleotide shuffled sequences !!! */
z_score_type = 0;
/* Init regression SVM for mean and standard deviation of MFE */
regression_svm_init(modelDir);
structure = (char *) space((unsigned) length+1);
structure_reverse = (char *) space((unsigned) length+1);
min_en = alifold(window, structure);
min_en_reverse = alifold(window_reverse, structure_reverse);
/* get z-score, mean mfe, average single mfe and sci*/
for (i=0;i<n_seq;i++)
{
singleStruc = space(strlen(window[i])+1);
woGapsSeq = space(strlen(window[i])+1);
singleStruc_reverse = space(strlen(window_reverse[i])+1);
woGapsSeq_reverse = space(strlen(window_reverse[i])+1);
j=0;
while (window[i][j])
{
window[i][j]=toupper(window[i][j]);
if (window[i][j]=='T') window[i][j]='U';
if (window[i][j]!='-')
{
woGapsSeq[strlen(woGapsSeq)]=window[i][j];
woGapsSeq[strlen(woGapsSeq)]='\0';
}
++j;
}
j=0;
while (window_reverse[i][j])
{
window_reverse[i][j]=toupper(window_reverse[i][j]);
if (window_reverse[i][j]=='T') window_reverse[i][j]='U';
if (window_reverse[i][j]!='-')
{
woGapsSeq_reverse[strlen(woGapsSeq_reverse)]=window_reverse[i][j];
woGapsSeq_reverse[strlen(woGapsSeq_reverse)]='\0';
}
++j;
}
singleMFE = fold(woGapsSeq, singleStruc);
singleZ=mfe_zscore(woGapsSeq,singleMFE,z_score_type,0,warningString_strand);
singleMFE_reverse = fold(woGapsSeq_reverse, singleStruc_reverse);
singleZ_reverse=mfe_zscore(woGapsSeq_reverse,singleMFE_reverse,z_score_type,0,warningString_strand);
sumZ+=singleZ;
sumMFE+=singleMFE;
sumZ_reverse+=singleZ_reverse;
sumMFE_reverse+=singleMFE_reverse;
free(woGapsSeq);
free(singleStruc);
free(woGapsSeq_reverse);
free(singleStruc_reverse);
}
{
int i; double s=0;
extern int eos_debug;
eos_debug=-1; /* shut off warnings about nonstandard pairs */
for (i=0; window[i]!=NULL; i++)
s += energy_of_struct(window[i], structure);
real_en = s/i;
}
z=sumZ/n_seq;
sci=min_en/(sumMFE/n_seq);
z_reverse=sumZ_reverse/n_seq;
sci_reverse=min_en_reverse/(sumMFE_reverse/n_seq);
*deltaSCI = sci - sci_reverse;
*deltaMeanMFE = sumMFE/n_seq - sumMFE_reverse/n_seq;
*deltaConsMFE = min_en - min_en_reverse;
*deltaZ = z - z_reverse;
/* Free svm regression model */
regression_svm_free();
return;
}
int main(int argc, char *argv[])
{
char *string;
char *structure=NULL;
char *cstruc=NULL;
char *ns_bases=NULL;
char *c;
int n_seq;
int i;
int length;
int sym;
int endgaps = 0;
int mis = 0;
double min_en;
double real_en;
double sfact = 1.07;
int pf = 0;
int istty;
char *AS[MAX_NUM_NAMES]; /* aligned sequences */
char *names[MAX_NUM_NAMES]; /* sequence names */
AjPSeqset seq = NULL;
AjPFile confile = NULL;
AjPFile alifile = NULL;
AjPFile paramfile = NULL;
AjPFile outf = NULL;
AjPFile essfile = NULL;
AjPFile dotfile = NULL;
AjPStr constring = NULL;
float eT = 0.;
AjBool eGU;
AjBool eclose;
AjBool lonely;
AjPStr ensbases = NULL;
AjBool etloop;
AjPStr eenergy = NULL;
char ewt = '\0';
float escale = 0.;
AjPStr edangles = NULL;
char edangle = '\0';
ajint len;
AjPSeq tseq = NULL;
AjPStr tname = NULL;
int circ = 0;
int doAlnPS = 0;
int doColor = 0;
embInitPV("vrnaalifoldpf",argc,argv,"VIENNA",VERSION);
constring = ajStrNew();
seq = ajAcdGetSeqset("sequence");
confile = ajAcdGetInfile("constraintfile");
paramfile = ajAcdGetInfile("paramfile");
eT = ajAcdGetFloat("temperature");
eGU = ajAcdGetBoolean("gu");
eclose = ajAcdGetBoolean("closegu");
lonely = ajAcdGetBoolean("lp");
ensbases = ajAcdGetString("nsbases");
etloop = ajAcdGetBoolean("tetraloop");
eenergy = ajAcdGetListSingle("energy");
escale = ajAcdGetFloat("scale");
edangles = ajAcdGetListSingle("dangles");
mis = !!ajAcdGetBoolean("most");
endgaps = !!ajAcdGetBoolean("endgaps");
nc_fact = (double) ajAcdGetFloat("nspenalty");
cv_fact = (double) ajAcdGetFloat("covariance");
outf = ajAcdGetOutfile("outfile");
essfile = ajAcdGetOutfile("ssoutfile");
alifile = ajAcdGetOutfile("alignoutfile");
circ = !!ajAcdGetBoolean("circular");
doColor = !!ajAcdGetBoolean("colour");
dotfile = ajAcdGetOutfile("dotoutfile");
do_backtrack = 1;
pf = 1;
string = NULL;
istty = 0;
dangles = 2;
temperature = (double) eT;
noGU = (eGU) ? 0 : 1;
no_closingGU = (eclose) ? 0 : 1;
noLonelyPairs = (lonely) ? 0 : 1;
ns_bases = (ajStrGetLen(ensbases)) ? MAJSTRGETPTR(ensbases) : NULL;
tetra_loop = !!etloop;
ewt = *ajStrGetPtr(eenergy);
if(ewt == '0')
energy_set = 0;
else if(ewt == '1')
energy_set = 1;
else if(ewt == '2')
energy_set = 2;
sfact = (double) escale;
edangle = *ajStrGetPtr(edangles);
if(edangle == '0')
dangles = 0;
else if(edangle == '1')
dangles = 1;
else if(edangle == '2')
dangles = 2;
else if(edangle == '3')
dangles = 3;
if(paramfile)
read_parameter_file(paramfile);
if (ns_bases != NULL)
{
nonstandards = space(33);
c=ns_bases;
i=sym=0;
if (*c=='-')
{
sym=1;
c++;
}
while (*c!='\0')
{
if (*c!=',')
{
nonstandards[i++]=*c++;
nonstandards[i++]=*c;
if ((sym)&&(*c!=*(c-1)))
{
nonstandards[i++]=*c;
nonstandards[i++]=*(c-1);
}
}
c++;
}
}
if(alifile)
doAlnPS = 1;
if(confile)
vienna_GetConstraints(confile,&constring);
n_seq = ajSeqsetGetSize(seq);
if(n_seq > MAX_NUM_NAMES - 1)
ajFatal("[e]RNAalifold is restricted to %d sequences\n",
MAX_NUM_NAMES - 1);
if (n_seq==0)
ajFatal("No sequences found");
for(i=0;i<n_seq;++i)
{
tseq = (AjPSeq) ajSeqsetGetseqSeq(seq,i);
ajSeqGapStandard(tseq, '-');
tname = (AjPStr) ajSeqsetGetseqNameS(seq,i);
len = ajSeqGetLen(tseq);
AS[i] = (char *) space(len+1);
names[i] = (char *) space(ajStrGetLen(tname)+1);
strcpy(AS[i],ajSeqGetSeqC(tseq));
strcpy(names[i],ajStrGetPtr(tname));
}
AS[n_seq] = NULL;
names[n_seq] = NULL;
if (endgaps)
for (i=0; i<n_seq; i++)
mark_endgaps(AS[i], '~');
length = (int) strlen(AS[0]);
structure = (char *) space((unsigned) length+1);
if(confile)
{
fold_constrained = 1;
strcpy(structure,ajStrGetPtr(constring));
}
if (circ && noLonelyPairs)
ajWarn(
"warning, depending on the origin of the circular sequence, "
"some structures may be missed when using -noLP\n"
"Try rotating your sequence a few times\n");
if (circ)
min_en = circalifold((const char **)AS, structure);
else
min_en = alifold(AS, structure);
{
int i;
double s=0;
extern int eos_debug;
eos_debug=-1; /* shut off warnings about nonstandard pairs */
for (i=0; AS[i]!=NULL; i++)
if (circ)
s += energy_of_circ_struct(AS[i], structure);
else
s += energy_of_struct(AS[i], structure);
real_en = s/i;
}
string = (mis) ?
consens_mis((const char **) AS) : consensus((const char **) AS);
ajFmtPrintF(outf,"%s\n%s", string, structure);
ajFmtPrintF(outf," (%6.2f = %6.2f + %6.2f) \n", min_en, real_en,
min_en-real_en );
if (length<=2500) {
char **A;
A = annote(structure, (const char**) AS);
if (doColor)
(void) PS_rna_plot_a(string, structure, essfile, A[0], A[1]);
else
(void) PS_rna_plot_a(string, structure, essfile, NULL, A[1]);
free(A[0]); free(A[1]);free(A);
} else
ajWarn("INFO: structure too long, not doing xy_plot\n");
if (doAlnPS)
PS_color_aln(structure, alifile, AS, names);
{ /* free mfe arrays but preserve base_pair for PS_dot_plot */
struct bond *bp;
bp = base_pair; base_pair = space(16);
free_alifold_arrays(); /* free's base_pair */
free_alipf_arrays();
base_pair = bp;
}
if (pf) {
double energy, kT;
pair_info *pi;
char * mfe_struc;
mfe_struc = strdup(structure);
kT = (temperature+273.15)*1.98717/1000.; /* in Kcal */
pf_scale = exp(-(sfact*min_en)/kT/length);
if (length>2000)
ajWarn("scaling factor %f\n", pf_scale);
/* init_alipf_fold(length); */
if (confile)
strncpy(structure, ajStrGetPtr(constring), length+1);
energy = (circ) ? alipf_circ_fold(AS, structure, &pi) : alipf_fold(AS, structure, &pi);
if (do_backtrack) {
ajFmtPrintF(outf,"%s", structure);
ajFmtPrintF(outf," [%6.2f]\n", energy);
}
if ((istty)||(!do_backtrack))
ajFmtPrintF(outf," free energy of ensemble = %6.2f kcal/mol\n", energy);
ajFmtPrintF(outf," frequency of mfe structure in ensemble %g\n",
exp((energy-min_en)/kT));
if (do_backtrack) {
FILE *aliout;
cpair *cp;
short *ptable; int k;
ptable = make_pair_table(mfe_struc);
ajFmtPrintF(outf,"\n# Alignment section\n\n");
aliout = ajFileGetFileptr(outf);
fprintf(aliout, "%d sequences; length of alignment %d\n",
n_seq, length);
fprintf(aliout, "alifold output\n");
for (k=0; pi[k].i>0; k++) {
pi[k].comp = (ptable[pi[k].i] == pi[k].j) ? 1:0;
print_pi(pi[k], aliout);
}
fprintf(aliout, "%s\n", structure);
free(ptable);
cp = make_color_pinfo(pi);
(void) PS_color_dot_plot(string, cp, dotfile);
free(cp);
free(mfe_struc);
free(pi);
}
}
if (cstruc!=NULL) free(cstruc);
free(base_pair);
(void) fflush(stdout);
free(string);
free(structure);
for (i=0; AS[i]; i++) {
free(AS[i]); free(names[i]);
}
ajSeqsetDel(&seq);
ajStrDel(&constring);
ajStrDel(&eenergy);
ajStrDel(&edangles);
ajStrDel(&ensbases);
ajFileClose(&confile);
ajFileClose(¶mfile);
ajFileClose(&outf);
ajFileClose(&essfile);
ajFileClose(&alifile);
ajFileClose(&dotfile);
embExit();
return 0;
}
int main(int argc, char *argv[])
{
char *string, *line;
char *structure=NULL, *cstruc=NULL;
char fname[53], ffname[60];
char *ParamFile=NULL;
char *ns_bases=NULL, *c;
char *Concfile;
int i, length, l, sym, r;
double min_en;
double kT, sfact=1.07;
int pf=0, istty;
int noconv=0;
int doT=0; /*compute dimere free energies etc.*/
int doC=0; /*toggle to compute concentrations*/
int doQ=0; /*toggle to compute prob of base being paired*/
int cofi=0; /*toggle concentrations stdin / file*/
struct plist *prAB;
struct plist *prAA; /*pair probabilities of AA dimer*/
struct plist *prBB;
struct plist *prA;
struct plist *prB;
struct plist *mfAB;
struct plist *mfAA; /*pair mfobabilities of AA dimer*/
struct plist *mfBB;
struct plist *mfA;
struct plist *mfB;
double *ConcAandB;
do_backtrack = 1;
string=NULL;
Concfile=NULL;
for (i=1; i<argc; i++) {
if (argv[i][0]=='-')
switch ( argv[i][1] )
{
case 'T':
if (argv[i][2]!='\0') usage();
if(i==argc-1) usage();
r=sscanf(argv[++i], "%lf", &temperature);
if (!r) usage();
break;
case 'p':
pf=1;
if (argv[i][2]!='\0')
(void) sscanf(argv[i]+2, "%d", &do_backtrack);
break;
case 'n':
if ( strcmp(argv[i], "-noGU")==0) noGU=1;
if ( strcmp(argv[i], "-noCloseGU")==0) no_closingGU=1;
if ( strcmp(argv[i], "-noLP")==0) noLonelyPairs=1;
if ( strcmp(argv[i], "-nsp") ==0) {
if (i==argc-1) usage();
ns_bases = argv[++i];
}
if ( strcmp(argv[i], "-noconv")==0) noconv=1;
break;
case '4':
tetra_loop=0;
break;
case 'e':
if(i==argc-1) usage();
r=sscanf(argv[++i],"%d", &energy_set);
if (!r) usage();
break;
case 'C':
fold_constrained=1;
break;
case 'S':
if(i==argc-1) usage();
r=sscanf(argv[++i],"%lf", &sfact);
if (!r) usage();
break;
case 'd':
dangles=0;
if (argv[i][2]!='\0') {
r=sscanf(argv[i]+2, "%d", &dangles);
if (r!=1) usage();
}
break;
case 'P':
if (i==argc-1) usage();
ParamFile = argv[++i];
break;
case 'a':
doT=1;
pf=1;
break;
case 'c':/*concentrations from stdin*/
doC=1;
doT=1;
pf=1;
break;
case 'f':/*concentrations in file*/
if (i==argc-1) usage();
Concfile = argv[++i];
doC=1;
cofi=1;
doT=1;
pf=1;
break;
case 'q':
pf=1;
doQ=1;
break;
default:
usage();
}
}
if (ParamFile != NULL)
read_parameter_file(ParamFile);
if (ns_bases != NULL) {
nonstandards = space(33);
c=ns_bases;
i=sym=0;
if (*c=='-') {
sym=1;
c++;
}
while (*c!='\0') {
if (*c!=',') {
nonstandards[i++]=*c++;
nonstandards[i++]=*c;
if ((sym)&&(*c!=*(c-1))) {
nonstandards[i++]=*c;
nonstandards[i++]=*(c-1);
}
}
c++;
}
}
istty = isatty(fileno(stdout))&&isatty(fileno(stdin));
if ((fold_constrained)&&(istty)) {
printf("Input constraints using the following notation:\n");
printf("| : paired with another base\n");
printf(". : no constraint at all\n");
printf("x : base must not pair\n");
printf("< : base i is paired with a base j<i\n");
printf("> : base i is paired with a base j>i\n");
printf("matching brackets ( ): base i pairs base j\n");
}
do { /* main loop: continue until end of file */
cut_point = -1;
if (istty) {
printf("\nInput sequence(s); @ to quit\n");
printf("Use '&' as spearator between 2 sequences that shall form a complex.\n");
printf("%s\n", scale);
}
fname[0]='\0';
if ((line = get_line(stdin))==NULL) break;
/* skip comment lines and get filenames */
while ((*line=='*')||(*line=='\0')||(*line=='>')) {
if (*line=='>')
(void) sscanf(line, ">%51s", fname);
printf("%s\n", line);
free(line);
if ((line = get_line(stdin))==NULL) break;
}
if ((line ==NULL) || (strcmp(line, "@") == 0)) break;
string = tokenize(line); /* frees line */
length = (int) strlen(string);
if (doC) {
FILE *fp;
if (cofi) { /* read from file */
fp = fopen(Concfile, "r");
if (fp==NULL) {
fprintf(stderr, "could not open concentration file %s", Concfile);
nrerror("\n");
}
ConcAandB = read_concentrations(fp);
fclose(fp);
} else {
printf("Please enter concentrations [mol/l]\n format: ConcA ConcB\n return to end\n");
ConcAandB = read_concentrations(stdin);
}
}
structure = (char *) space((unsigned) length+1);
if (fold_constrained) {
cstruc = tokenize(get_line(stdin));
if (cstruc!=NULL)
strncpy(structure, cstruc, length);
else
fprintf(stderr, "constraints missing\n");
}
for (l = 0; l < length; l++) {
string[l] = toupper(string[l]);
if (!noconv && string[l] == 'T') string[l] = 'U';
}
if (istty) {
if (cut_point == -1)
printf("length = %d\n", length);
else
printf("length1 = %d\nlength2 = %d\n",
cut_point-1, length-cut_point+1);
}
/*compute mfe of AB dimer*/
min_en = cofold(string, structure);
mfAB=(struct plist *) space(sizeof(struct plist) * (length+1));
mfAB=get_mfe_plist(mfAB);
if (cut_point == -1) printf("%s\n%s", string, structure); /*no cofold*/
else {
char *pstring, *pstruct;
pstring = costring(string);
pstruct = costring(structure);
printf("%s\n%s", pstring, pstruct);
free(pstring);
free(pstruct);
}
if (istty)
printf("\n minimum free energy = %6.2f kcal/mol\n", min_en);
else
printf(" (%6.2f)\n", min_en);
(void) fflush(stdout);
if (fname[0]!='\0') {
strcpy(ffname, fname);
strcat(ffname, "_ss.ps");
} else {
strcpy(ffname, "rna.ps");
}
if (length<2000)
(void) PS_rna_plot(string, structure, ffname);
else {
fprintf(stderr,"INFO: structure too long, not doing xy_plot\n");
free_co_arrays();
}
/*compute partition function*/
if (pf) {
cofoldF AB, AA, BB;
if (dangles==1) {
dangles=2; /* recompute with dangles as in pf_fold() */
min_en = energy_of_struct(string, structure);
dangles=1;
}
kT = (temperature+273.15)*1.98717/1000.; /* in Kcal */
pf_scale = exp(-(sfact*min_en)/kT/length);
if (length>2000) fprintf(stderr, "scaling factor %f\n", pf_scale);
init_co_pf_fold(length);
if (cstruc!=NULL)
strncpy(structure, cstruc, length+1);
AB = co_pf_fold(string, structure);
if (do_backtrack) {
char *costruc;
costruc = (char *) space(sizeof(char)*(strlen(structure)+2));
if (cut_point<0) printf("%s", structure);
else {
strncpy(costruc, structure, cut_point-1);
strcat(costruc, "&");
strcat(costruc, structure+cut_point-1);
printf("%s", costruc);
}
if (!istty) printf(" [%6.2f]\n", AB.FAB);
else printf("\n");/*8.6.04*/
}
if ((istty)||(!do_backtrack))
printf(" free energy of ensemble = %6.2f kcal/mol\n", AB.FAB);
printf(" frequency of mfe structure in ensemble %g",
exp((AB.FAB-min_en)/kT));
printf(" , delta G binding=%6.2f\n", AB.FcAB - AB.FA - AB.FB);
prAB=(struct plist *) space(sizeof(struct plist) * (2*length));
prAB=get_plist(prAB, length,0.00001);
/* if (doQ) make_probsum(length,fname); */ /*compute prob of base paired*/
/* free_co_arrays(); */
if (doT) { /* cofold of all dimers, monomers */
int Blength, Alength;
char *Astring, *Bstring;
char *Newstring;
char Newname[30];
char comment[80];
if (cut_point<0) {
printf("Sorry, i cannot do that with only one molecule, please give me two or leave it\n");
free(mfAB);
free(prAB);
continue;
}
if (dangles==1) dangles=2;
Alength=cut_point-1; /*length of first molecule*/
Blength=length-cut_point+1; /*length of 2nd molecule*/
Astring=(char *)space(sizeof(char)*(Alength+1));/*Sequence of first molecule*/
Bstring=(char *)space(sizeof(char)*(Blength+1));/*Sequence of second molecule*/
strncat(Astring,string,Alength);
strncat(Bstring,string+Alength,Blength);
/* compute AA dimer */
prAA=(struct plist *) space(sizeof(struct plist) * (4*Alength));
mfAA=(struct plist *) space(sizeof(struct plist) * (Alength+1));
AA=do_partfunc(Astring, Alength, 2, &prAA, &mfAA);
/* compute BB dimer */
prBB=(struct plist *) space(sizeof(struct plist) * (4*Blength));
mfBB=(struct plist *) space(sizeof(struct plist) * (Blength+1));
BB=do_partfunc(Bstring, Blength, 2, &prBB, &mfBB);
/*free_co_pf_arrays();*/
/* compute A monomer */
prA=(struct plist *) space(sizeof(struct plist) * (2*Alength));
mfA=(struct plist *) space(sizeof(struct plist) * (Alength+1));
do_partfunc(Astring, Alength, 1, &prA, &mfA);
/* compute B monomer */
prB=(struct plist *) space(sizeof(struct plist) * (2*Blength));
mfB=(struct plist *) space(sizeof(struct plist) * (Blength+1));
do_partfunc(Bstring, Blength, 1, &prB, &mfB);
compute_probabilities(AB.F0AB, AB.FA, AB.FB, prAB, prA, prB, Alength);
compute_probabilities(AA.F0AB, AA.FA, AA.FA, prAA, prA, prA, Alength);
compute_probabilities(BB.F0AB, BB.FA, BB.FA, prBB, prA, prB, Blength);
printf("Free Energies:\nAB\t\tAA\t\tBB\t\tA\t\tB\n%.6f\t%6f\t%6f\t%6f\t%6f\n",
AB.FcAB, AA.FcAB, BB.FcAB, AB.FA, AB.FB);
if (doC) {
do_concentrations(AB.FcAB, AA.FcAB, BB.FcAB, AB.FA, AB.FB, ConcAandB);
free(ConcAandB);/*freeen*/
}
if (fname[0]!='\0') {
strcpy(ffname, fname);
strcat(ffname, "_dp5.ps");
} else strcpy(ffname, "dot5.ps");
/*output of the 5 dot plots*/
/*AB dot_plot*/
/*write Free Energy into comment*/
sprintf(comment,"\n%%Heterodimer AB FreeEnergy= %.9f\n", AB.FcAB);
/*reset cut_point*/
cut_point=Alength+1;
/*write New name*/
strcpy(Newname,"AB");
strcat(Newname,ffname);
(void)PS_dot_plot_list(string, Newname, prAB, mfAB, comment);
/*AA dot_plot*/
sprintf(comment,"\n%%Homodimer AA FreeEnergy= %.9f\n",AA.FcAB);
/*write New name*/
strcpy(Newname,"AA");
strcat(Newname,ffname);
/*write AA sequence*/
Newstring=(char*)space((2*Alength+1)*sizeof(char));
strcpy(Newstring,Astring);
strcat(Newstring,Astring);
(void)PS_dot_plot_list(Newstring, Newname, prAA, mfAA, comment);
free(Newstring);
/*BB dot_plot*/
sprintf(comment,"\n%%Homodimer BB FreeEnergy= %.9f\n",BB.FcAB);
/*write New name*/
strcpy(Newname,"BB");
strcat(Newname,ffname);
/*write BB sequence*/
Newstring=(char*)space((2*Blength+1)*sizeof(char));
strcpy(Newstring,Bstring);
strcat(Newstring,Bstring);
/*reset cut_point*/
cut_point=Blength+1;
(void)PS_dot_plot_list(Newstring, Newname, prBB, mfBB, comment);
free(Newstring);
/*A dot plot*/
/*reset cut_point*/
cut_point=-1;
sprintf(comment,"\n%%Monomer A FreeEnergy= %.9f\n",AB.FA);
/*write New name*/
strcpy(Newname,"A");
strcat(Newname,ffname);
/*write BB sequence*/
(void)PS_dot_plot_list(Astring, Newname, prA, mfA, comment);
/*B monomer dot plot*/
sprintf(comment,"\n%%Monomer B FreeEnergy= %.9f\n",AB.FB);
/*write New name*/
strcpy(Newname,"B");
strcat(Newname,ffname);
/*write BB sequence*/
(void)PS_dot_plot_list(Bstring, Newname, prB, mfB, comment);
free(Astring);
free(Bstring);
free(prAB);
free(prAA);
free(prBB);
free(prA);
free(prB);
free(mfAB);
free(mfAA);
free(mfBB);
free(mfA);
free(mfB);
} /*end if(doT)*/
}/*end if(pf)*/
if (do_backtrack) {
if (fname[0]!='\0') {
strcpy(ffname, fname);
strcat(ffname, "_dp.ps");
} else strcpy(ffname, "dot.ps");
if (!doT) {
if (pf) {
(void) PS_dot_plot_list(string, ffname, prAB, mfAB, "doof");
free(prAB);
}
free(mfAB);
}
}
if (!doT) free_co_pf_arrays();
if (cstruc!=NULL) free(cstruc);
(void) fflush(stdout);
free(string);
free(structure);
} while (1);
return 0;
}
static int comp_struc(const void *A, const void *B) {
int aE, bE;
aE = (int)(100 * energy_of_struct(GAV.farbe_full, ((char **)A)[0]));
bE = (int)(100 * energy_of_struct(GAV.farbe_full, ((char **)B)[0]));
return (aE-bE);
}
/*--------------------------------------------------------------------------*/
int main(int argc, char *argv[])
{
char *string1=NULL, *string2=NULL, *temp, *line;
char *structure=NULL, *cstruc=NULL;
char fname[53], my_contrib[10], *up_out;
char *ParamFile=NULL;
char *ns_bases=NULL, *c;
int i, length1,length2,length, l, sym, r;
double energy, min_en;
double kT, sfact=1.07;
int pf, istty;
int noconv=0;
double Zu, Zup;
/* variables for output */
pu_contrib *unstr_out, *unstr_short;
FLT_OR_DBL **inter_out;
char *title;
/* commandline parameters */
int w; /* length of region of interaction */
int incr3; /* add x unpaired bases after 3'end of short RNA*/
int incr5; /* add x unpaired bases after 5'end of short RNA*/
int unstr; /* length of unpaired region for output*/
int upmode; /* output mode for pf_unpaired and pf_up()*/
upmode = 0;
unstr = 4;
incr3=0;
incr5=0;
w=25;
do_backtrack = 1;
pf=1; /* partition function has to be calculated */
length1=length2=0;
up_out=NULL;
title=NULL;
unstr_out=NULL;
inter_out=NULL;
my_contrib[0] = 'S';
my_contrib[1] = '\0';
for (i=1; i<argc; i++) {
if (argv[i][0]=='-')
switch ( argv[i][1] )
{
case 'T': if (argv[i][2]!='\0') usage();
if(i==argc-1) usage();
r=sscanf(argv[++i], "%lf", &temperature);
if (!r) usage();
break;
case 'w':
/* -w maximal length of unstructured region */
if(i==argc-1) usage();
r=sscanf(argv[++i],"%d", &w);
if (!r) usage();
break;
case 'n':
if ( strcmp(argv[i], "-noGU")==0) noGU=1;
if ( strcmp(argv[i], "-noCloseGU")==0) no_closingGU=1;
if ( strcmp(argv[i], "-noLP")==0) noLonelyPairs=1;
if ( strcmp(argv[i], "-nsp") ==0) {
if (i==argc-1) usage();
ns_bases = argv[++i];
}
if ( strcmp(argv[i], "-noconv")==0) noconv=1;
break;
case '4':
tetra_loop=0;
break;
case 'e':
if(i==argc-1) usage();
r=sscanf(argv[++i],"%d", &energy_set);
if (!r) usage();
break;
case 'C':
fold_constrained=1;
break;
case 'S':
if(i==argc-1) usage();
r=sscanf(argv[++i],"%lf", &sfact);
if (!r) usage();
break;
case 'd': dangles=0;
if (argv[i][2]!='\0') {
r=sscanf(argv[i]+2, "%d", &dangles);
if (r!=1) usage();
}
break;
case 'o': upmode=1;
/* output mode 0: non, 1:only pr_unpaired, 2: pr_unpaired + pr_up */
if (argv[i][2]!='\0') {
r=sscanf(argv[i]+2, "%d", &upmode);
if (r!=1) usage();
}
break;
case 'u':
/* -u length of unstructured region in pr_unpaired output
makes only sense in combination with -o1 or -o2 */
if(i==argc-1) usage();
r=sscanf(argv[++i],"%d", &unstr);
if (!r) usage();
break;
/* incr5 and incr3 are only for the longer (target) sequence */
/* increments w (length of the unpaired region) to incr5+w+incr3*/
/* the longer sequence is given in 5'(= position 1) to */
/* 3' (=position n) direction */
/* incr5 adds incr5 residues to the 5' end of w */
case '5':
if(i==argc-1) usage();
r=sscanf(argv[++i],"%d", &incr5);
if (!r) usage();
break;
/* incr3 adds incr3 residues to the 3' end of w */
case '3':
if(i==argc-1) usage();
r=sscanf(argv[++i],"%d", &incr3);
if (!r) usage();
break;
case 'P':
if (i==argc-1) usage();
ParamFile = argv[++i];
break;
case 'x':
if(i==argc-1) usage();
r=sscanf(argv[++i], "%s", my_contrib);
if (!r) usage();
break;
default: usage();
}
}
if (ParamFile != NULL)
read_parameter_file(ParamFile);
if (ns_bases != NULL) {
nonstandards = space(33);
c=ns_bases;
i=sym=0;
if (*c=='-') {
sym=1; c++;
}
while (*c!='\0') {
if (*c!=',') {
nonstandards[i++]=*c++;
nonstandards[i++]=*c;
if ((sym)&&(*c!=*(c-1))) {
nonstandards[i++]=*c;
nonstandards[i++]=*(c-1);
}
}
c++;
}
}
istty = isatty(fileno(stdout))&&isatty(fileno(stdin));
if ((fold_constrained)&&(istty)) {
printf("Input constraints using the following notation:\n");
printf("| : paired with another base\n");
printf(". : no constraint at all\n");
printf("x : base must not pair\n");
printf("< : base i is paired with a base j<i\n");
printf("> : base i is paired with a base j>i\n");
printf("matching brackets ( ): base i pairs base j\n");
}
do { /* main loop: continue until end of file */
cut_point=-1;
if (istty) {
printf("\nInput string (upper or lower case); @ to quit\n");
printf("Use '&' to connect 2 sequences that shall form a complex.\n");
printf("%s%s\n", scale1, scale2);
}
fname[0]='\0';
if ((line = get_line(stdin))==NULL) break;
/* skip comment lines and get filenames */
while ((*line=='*')||(*line=='\0')||(*line=='>')) {
if (*line=='>')
(void) sscanf(line, ">%51s", fname);
free(line);
if ((line = get_line(stdin))==NULL) break;
}
if ((line == NULL) || (strcmp(line, "@") == 0)) break;
tokenize(line,&string1,&string2);
if(upmode != 0){
if(cut_point == -1 && upmode == 2) {
nrerror("only one sequence - can not cofold one sequence!");
}
} else {
if(cut_point == -1){
upmode=1;
} else {
upmode=2;
}
}
if(string1 != NULL)
length1 = (int) strlen(string1);
if(string2 != NULL)
length2 = (int) strlen(string2);
else
length2=0;
/* write longer seq in string1 and and shorter one in string2 */
if(length1 < length2)
{
length=length1; length1=length2; length2=length;
temp=(char *) space(strlen(string1)+1);
(void) sscanf(string1,"%s",temp);
string1 = (char *) xrealloc (string1,sizeof(char)*length1+1);
(void) sscanf(string2,"%s",string1);
string2 = (char *) xrealloc(string2,sizeof(char)*length2+1);
(void) sscanf(temp,"%s",string2);
free(temp);
}
structure = (char *) space((unsigned) length1+1);
if (fold_constrained) {
cstruc = get_line(stdin);
if (cstruc!=NULL)
strncpy(structure, cstruc, length1);
else
fprintf(stderr, "constraints missing\n");
}
for (l = 0; l < length1; l++) {
string1[l] = toupper(string1[l]);
if (!noconv && string1[l] == 'T') string1[l] = 'U';
}
for (l = 0; l < length2; l++) {
string2[l] = toupper(string2[l]);
if (!noconv && string2[l] == 'T') string2[l] = 'U';
}
if (istty)
printf("length1 = %d\n", length1);
/* initialize_fold(length); */
update_fold_params();
printf("\n%s", string1);
min_en = fold(string1, structure);
if (istty)
{
printf("\n minimum free energy = %6.2f kcal/mol\n", min_en);
}
else
printf(" (%6.2f)\n", min_en);
(void) fflush(stdout);
/* parse cml parameters for the filename*/
if(upmode > 0) {
char wuadd[10];
up_out = (char*) space(sizeof(char)*53);
/* create the name of the output file */
if(fname[0]!='\0' && up_out[0] =='\0' ){
if(strlen(fname)< 30){
strcpy(up_out, fname);
} else {
strncpy(up_out, fname,30);
}
}
else if(fname[0]=='\0' && up_out[0] == '\0'){
char defaultn[10] = "RNA";
sprintf(up_out,"%s",defaultn);
}
sprintf(wuadd,"%d",w);
strcat(up_out, "_w");
strcat(up_out, wuadd);
strcat(up_out, "u");
sprintf(wuadd,"%d",unstr);
strcat(up_out, wuadd);
strcat(up_out, "_up.out");
printf("RNAup output in file: %s\n",up_out);
/* create the title for the output file */
if (title == NULL) {
char wuadd[10];
title = (char*) space(sizeof(char)*60);
if(fname[0]!='\0'){
if(strlen(fname)< 30){
strcpy(title, fname);
} else {
strncpy(title, fname,30);
}
}
else if (fname[0]=='\0'){
char defaultn[10]= "RNAup";
sprintf(title,"%s",defaultn);
}
sprintf(wuadd,"%d",unstr);
strcat(title," u=");
strcat(title, wuadd);
sprintf(wuadd,"%d",w);
strcat(title," w=");
strcat(title, wuadd);
sprintf(wuadd,"%d",length1);
strcat(title," n=");
strcat(title, wuadd);
}
} else {
nrerror("no output format given: use [-o[1|2]] to select output format");
}
if (pf) {
if (dangles==1) {
dangles=2; /* recompute with dangles as in pf_fold() */
min_en = energy_of_struct(string1, structure);
dangles=1;
}
kT = (temperature+273.15)*1.98717/1000.; /* in Kcal */
if(upmode != 0){
int wplus;
wplus=w+incr3+incr5;
/* calculate prob. unstructured for the shorter seq */
if(upmode == 3) {
min_en = fold(string2, structure);
pf_scale = exp(-(sfact*min_en)/kT/length2);
if (length2>2000) fprintf(stderr, "scaling factor %f\n", pf_scale);
init_pf_fold(length2);
if (cstruc!=NULL)
strncpy(structure, cstruc, length2+1);
energy = pf_fold(string2, structure);
if(wplus > length2){ wplus = length2;} /* for the shorter seq */
unstr_short = pf_unstru(string2, structure, wplus);
free_pf_unstru();
free_pf_arrays(); /* for arrays for pf_fold(...) */
}
/* calculate prob. unstructured for the longer seq */
wplus=w+incr3+incr5;
min_en = fold(string1, structure);
pf_scale = exp(-(sfact*min_en)/kT/length1);
if (length1>2000) fprintf(stderr, "scaling factor %f\n", pf_scale);
init_pf_fold(length1);
if (cstruc!=NULL)
strncpy(structure, cstruc, length1+1);
energy = pf_fold(string1, structure);
unstr_out = pf_unstru(string1, structure, wplus);
free_pf_unstru();
free_pf_arrays(); /* for arrays for pf_fold(...) */
/* calculate the interaction between the two sequences */
if(upmode > 1 && cut_point > -1){
inter_out = pf_interact(string1,string2,unstr_out,w, incr3, incr5);
if(Up_plot(unstr_out,inter_out,length1,up_out,unstr,my_contrib)==0){
nrerror("Up_plot: no output values assigned");
}
} else if(cut_point == -1 && upmode > 1) { /* no second seq given */
nrerror("only one sequence given - cannot cofold one sequence!");
} else { /* plot only the results for prob unstructured */
if(Up_plot(unstr_out,NULL,length1,up_out,unstr,my_contrib)==0){
nrerror("Up_plot: no output values assigned");
}
}
} else {
nrerror("no output format given: use [-o[1|2]] to select output format");
}
if (do_backtrack) {
printf("%s", structure);
if (!istty) printf(" [%6.2f]\n", energy);
else printf("\n");
}
if ((istty)||(!do_backtrack))
printf(" free energy of ensemble = %6.2f kcal/mol\n", energy);
energy = pf_fold(string1, structure);
printf(" frequency of mfe structure in ensemble %g; "
"ensemble diversity %-6.2f\n", exp((energy-min_en)/kT),
mean_bp_dist(length1));
free_pf_arrays();
}
if (cstruc!=NULL) free(cstruc);
(void) fflush(stdout);
if (string1!=NULL) free(string1);
if (string2!=NULL) free(string2);
free(structure);
if(up_out != NULL) free(up_out);
up_out=NULL;
if(title != NULL) free(title);
title=NULL;
if(upmode == 1) free_pf_two(unstr_out,NULL);
if(upmode > 1) free_pf_two(unstr_out,inter_out);
if(upmode == 3)free_pf_two(unstr_short,NULL);
free_arrays(); /* for arrays for fold(...) */
} while (1);
return 0;
}
