float fold(const char *string, char *structure) {
int i, length, energy, bonus=0, bonus_cnt=0;
circ = 0;
length = (int) strlen(string);
if (length>init_length) initialize_fold(length);
if (fabs(P->temperature - temperature)>1e-6) update_fold_params();
encode_seq(string);
BP = (int *)space(sizeof(int)*(length+2));
make_ptypes(S, structure);
energy = fill_arrays(string);
backtrack(string, 0);
#ifdef PAREN
parenthesis_structure(structure, length);
#else
letter_structure(structure, length);
#endif
/* check constraints */
for(i=1;i<=length;i++) {
if((BP[i]<0)&&(BP[i]>-4)) {
bonus_cnt++;
if((BP[i]==-3)&&(structure[i-1]==')')) bonus++;
if((BP[i]==-2)&&(structure[i-1]=='(')) bonus++;
if((BP[i]==-1)&&(structure[i-1]!='.')) bonus++;
}
if(BP[i]>i) {
int l;
bonus_cnt++;
for(l=1; l<=base_pair[0].i; l++)
if((i==base_pair[l].i)&&(BP[i]==base_pair[l].j)) bonus++;
}
}
if (bonus_cnt>bonus) fprintf(stderr,"\ncould not enforce all constraints\n");
bonus*=BONUS;
free(S); free(S1); free(BP);
energy += bonus; /*remove bonus energies from result */
if (backtrack_type=='C')
return (float) c[indx[length]+1]/100.;
else if (backtrack_type=='M')
return (float) fML[indx[length]+1]/100.;
else
return (float) energy/100.;
}
PRIVATE void heat_capacity(char *string, float T_min, float T_max,
float h, int m)
{
int length, i;
char *structure;
float hc, kT, min_en;
length = (int) strlen(string);
do_backtrack = 0;
temperature = T_min -m*h;
initialize_fold(length);
structure = (char *) space((unsigned) length+1);
min_en = fold(string, structure);
free(structure); free_arrays();
kT = (temperature+K0)*GASCONST/1000; /* in kcal */
pf_scale = exp(-(1.07*min_en)/kT/length );
init_pf_fold(length);
for (i=0; i<2*m+1; i++) {
F[i] = pf_fold(string, NULL); /* T_min -2h */
temperature += h;
kT = (temperature+K0)*GASCONST/1000;
pf_scale=exp(-(F[i]/length +h*0.00727)/kT); /* try to extrapolate F */
update_pf_params(length);
}
while (temperature <= (T_max+m*h+h)) {
hc = - ddiff(F,h,m)* (temperature +K0 - m*h -h);
printf("%g %g\n", (temperature-m*h-h), hc);
for (i=0; i<2*m; i++)
F[i] = F[i+1];
F[2*m] = pf_fold(string, NULL);
temperature += h;
kT = (temperature+K0)*GASCONST/1000;
pf_scale=exp(-(F[i]/length +h*0.00727)/kT);
update_pf_params(length);
}
free_pf_arrays();
}
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;
}
}
}
int main(int argc, char *argv[])
{
char *start, *structure, *rstart, *str2, *line;
char *ParamFile=NULL;
int i,j, length, l, hd;
double energy=0., kT;
int pf, mfe, istty;
int repeat, found;
do_backtrack = 0;
pf = 0;
mfe = 1;
repeat = 0;
init_rand();
for (i=1; i<argc; i++) {
if (argv[i][0]=='-')
switch ( argv[i][1] )
{
case 'a':
symbolset = argv[++i];
/* symbolset should only have uppercase characters */
for (l = 0; l < (int)strlen(symbolset); l++)
symbolset[l] = toupper(symbolset[l]);
break;
case 'T':
if (argv[i][2]!='\0') usage();
if (sscanf(argv[++i], "%lf", &temperature)==0)
usage();
break;
case 'F':
mfe = 0;
pf = 0;
for(j=2; j<(int)strlen(argv[i]); j++) {
switch( argv[i][j] ) {
case 'm' :
mfe = 1;
break;
case 'p' :
pf = 1; /* old version had dangles=0 here */
break;
default :
usage();
}
}
break;
case 'R':
repeat = REPEAT_DEFAULT;
if(++i<argc)
if (sscanf(argv[i], "%d", &repeat)==0)
usage();
break;
case 'n':
if (strcmp(argv[i], "-noGU" )==0) noGU=1;
else if (strcmp(argv[i], "-noLP" )==0) noLonelyPairs=1;
else usage();
break;
case '4':
tetra_loop=0;
break;
case 'e':
if (sscanf(argv[++i],"%d", &energy_set)==0)
usage();
break;
case 'd':
dangles=0;
if (argv[i][2]!='\0')
if (sscanf(argv[i]+2, "%d", &dangles)==0)
usage();
break;
case 'f': /* when to stop RNAfold -p */
if (sscanf(argv[++i],"%f", &final_cost)==0)
usage();
break;
case 'P':
if (++i<argc)
ParamFile = argv[i];
else
usage();
break;
case 'v':
inv_verbose = 1;
break;
default:
usage();
}
}
kT = (temperature+273.15)*1.98717/1000.0;
istty = (isatty(fileno(stdout))&&isatty(fileno(stdin)));
if (ParamFile!=NULL)
read_parameter_file(ParamFile);
give_up = (repeat<0);
do {
if (istty) {
printf("\nInput structure & start string"
" (lower case letters for const positions)\n"
" @ to quit, and 0 for random start string\n");
printf("%s\n", scale);
}
if ((line = get_line(stdin))==NULL) break;
/* read structure, skipping over comment lines */
while ((*line=='*')||(*line=='\0')||(*line=='>')) {
printf("%s\n", line);
free(line);
if ((line = get_line(stdin))==NULL) break;
}
/* stop at eof or '@' */
if (line==NULL) break;
if (strcmp(line, "@") == 0) {
free(line);
break;
}
structure = (char *) space(strlen(line)+1);
(void) sscanf(line,"%s",structure); /* scanf gets rid of trailing junk */
free(line);
length = (int) strlen(structure);
str2 = (char *) space((unsigned)length+1);
if ((line = get_line(stdin))!=NULL)
if (strcmp(line, "@") == 0) {
free(line);
break;
}
start = (char *) space((unsigned) length+1);
if (line !=NULL) {
(void) strncpy(start, line, length);
free(line);
}
if (istty) printf("length = %d\n", length);
if (repeat!=0) found = (repeat>0)? repeat : (-repeat);
else found = 1;
initialize_fold(length);
rstart = (char *) space((unsigned)length+1);
while(found>0) {
char *string;
string = (char *) space((unsigned)length+1);
strcpy(string, start);
for (i=0; i<length; i++) {
/* lower case characters are kept fixed, any other character
not in symbolset is replaced by a random character */
if (islower(string[i])) continue;
if (string[i]=='\0' || (strchr(symbolset,string[i])==NULL))
string[i]=symbolset[int_urn(0,strlen(symbolset)-1)];
}
strcpy(rstart, string); /* remember start string */
if (mfe) {
energy = inverse_fold(string, structure);
if( (repeat>=0) || (energy<=0.0) ) {
found--;
hd = hamming(rstart, string);
printf("%s %3d", string, hd);
if (energy>0) { /* no solution found */
printf(" d= %g\n", energy);
if(istty) {
energy = fold(string,str2);
printf("%s\n", str2);
}
} else printf("\n");
}
}
if (pf) {
if (!(mfe && give_up && (energy>0))) {
/* unless we gave up in the mfe part */
double prob, min_en, sfact=1.07;
/* get a reasonable pf_scale */
min_en = fold(string,str2);
pf_scale = exp(-(sfact*min_en)/kT/length);
init_pf_fold(length);
energy = inverse_pf_fold(string, structure);
prob = exp(-energy/kT);
hd = hamming(rstart, string);
printf("%s %3d (%g)\n", string, hd, prob);
free_pf_arrays();
}
if (!mfe) found--;
}
(void) fflush(stdout);
free(string);
}
free(rstart);
free_arrays();
free(structure);
free(str2);
free(start);
(void) fflush(stdout);
} while (1);
return 0;
}
int main(int argc, char *argv[])
{
char *start;
char *structure;
char *rstart;
char *str2;
char *line;
int i;
int length;
int l;
int hd;
double energy = 0.;
double kT;
int pf = 0;
int mfe = 0;
int istty;
int repeat;
int found;
AjPFile inf = NULL;
AjPSeq seq = NULL;
AjPFile paramfile = NULL;
AjPFile outf = NULL;
float eT = 0.;
AjBool eGU;
AjBool eclose;
AjBool lonely;
AjBool etloop;
AjPStr eenergy = NULL;
char ewt = '\0';
AjPStr edangles = NULL;
AjPStr method = NULL;
AjPStr ealpha = NULL;
AjBool showfails = ajFalse;
AjBool succeed = ajFalse;
char edangle = '\0';
ajint len;
FILE *fp;
embInitPV("vrnainverse",argc,argv,"VIENNA",VERSION);
inf = ajAcdGetInfile("structuresfile");
seq = ajAcdGetSeq("sequence");
paramfile = ajAcdGetInfile("paramfile");
eT = ajAcdGetFloat("temperature");
eGU = ajAcdGetBoolean("gu");
eclose = ajAcdGetBoolean("closegu");
lonely = ajAcdGetBoolean("lp");
etloop = ajAcdGetBoolean("tetraloop");
eenergy = ajAcdGetListSingle("energy");
edangles = ajAcdGetListSingle("dangles");
method = ajAcdGetListSingle("folding");
ealpha = ajAcdGetString("alphabet");
final_cost = ajAcdGetFloat("final");
repeat = ajAcdGetInt("repeats");
showfails = ajAcdGetBoolean("showfails");
succeed = ajAcdGetBoolean("succeed");
outf = ajAcdGetOutfile("outfile");
do_backtrack = 0;
structure = NULL;
istty = 0;
temperature = (double) eT;
noGU = (eGU) ? 0 : 1;
no_closingGU = (eclose) ? 0 : 1;
noLonelyPairs = (lonely) ? 0 : 1;
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;
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(ajStrMatchC(method,"mp"))
{
mfe = 1;
pf = 1;
}
else if(ajStrMatchC(method,"m"))
{
mfe = 1;
pf = 0;
}
else if(ajStrMatchC(method,"p"))
{
mfe = 0;
pf = 1;
}
len = ajStrGetLen(ealpha);
symbolset = (char *) space(len + 1);
strcpy(symbolset, ajStrGetPtr(ealpha));
for (l = 0; l < len; l++)
symbolset[l] = toupper(symbolset[l]);
inv_verbose = !!showfails;
fp = ajFileGetFileptr(inf);
init_rand();
kT = (temperature+273.15)*1.98717/1000.0;
istty = (isatty(fileno(stdout))&&isatty(fileno(stdin)));
if (paramfile)
read_parameter_file(paramfile);
give_up = succeed;
do {
if ((line = get_line(fp))==NULL) break;
/* read structure, skipping over comment lines */
while ((*line=='*')||(*line=='\0')||(*line=='>'))
{
free(line);
if ((line = get_line(fp))==NULL)
break;
}
/* stop at eof or '@' */
if (line==NULL) break;
if (strcmp(line, "@") == 0)
{
free(line);
break;
}
structure = (char *) space(strlen(line)+1);
/* scanf gets rid of trailing junk */
(void) sscanf(line,"%s",structure);
free(line);
length = (int) strlen(structure);
str2 = (char *) space((unsigned)length+1);
/* now look for a sequence to match the structure */
/*
if ((line = get_line(fp))!=NULL)
if (strcmp(line, "@") == 0)
{
free(line);
break;
}
*/
start = (char *) space((unsigned) length+1);
if(seq)
(void) strncpy(start, ajSeqGetSeqC(seq), length);
if (repeat!=0)
found = repeat;
else
found = 1;
initialize_fold(length);
rstart = (char *) space((unsigned)length+1);
while(found>0)
{
char *string;
string = (char *) space((unsigned)length+1);
strcpy(string, start);
for (i=0; i<length; i++)
{
/* lower case characters are kept fixed, any other character
not in symbolset is replaced by a random character */
if (islower(string[i]))
continue;
if (string[i]=='\0' || (strchr(symbolset,string[i])==NULL))
string[i]=symbolset[int_urn(0,strlen(symbolset)-1)];
}
strcpy(rstart, string); /* remember start string */
if (mfe)
{
energy = inverse_fold(string, structure);
if( (!succeed) || (energy<=0.0) ) {
found--;
hd = hamming(rstart, string);
ajFmtPrintF(outf,"%s %3d", string, hd);
if (energy>0)
{ /* no solution found */
ajFmtPrintF(outf," d = %f\n", energy);
}
else
ajFmtPrintF(outf,"\n");
}
}
if (pf)
{
if (!(mfe && give_up && (energy>0)))
{
/* unless we gave up in the mfe part */
double prob, min_en, sfact=1.07;
/* get a reasonable pf_scale */
min_en = fold(string,str2);
pf_scale = exp(-(sfact*min_en)/kT/length);
init_pf_fold(length);
energy = inverse_pf_fold(string, structure);
prob = exp(-energy/kT);
hd = hamming(rstart, string);
ajFmtPrintF(outf,"%s %3d (%f)\n", string, hd, prob);
free_pf_arrays();
}
if (!mfe)
found--;
}
free(string);
}
free(rstart);
free_arrays();
free(structure);
free(str2);
free(start);
} while (1);
ajSeqDel(&seq);
ajStrDel(&eenergy);
ajStrDel(&edangles);
ajStrDel(&method);
ajStrDel(&ealpha);
ajFileClose(&inf);
ajFileClose(¶mfile);
ajFileClose(&outf);
AJFREE(symbolset);
embExit();
return 0;
}
