void Make_Tree_4_Pars(arbre *tree, allseq *alldata, int n_site)
{
int i;
tree->site_pars = (int *)mCalloc(tree->n_pattern, sizeof(int));
tree->step_mat = (int *)mCalloc(tree->mod->ns * tree->mod->ns, sizeof(int));
For(i,2*tree->n_otu-3) Make_Edge_Pars(tree->t_edges[i],tree);
Init_Ui_Tips(tree);
Init_P_Pars_Tips(tree); /* Must be called after Init_Ui_Tips is called */
Get_Step_Mat(tree);
}
char *str_replace (const char *s, char ch, const char *repl)
{
int count = 0;
const char *t;
for (t=s; *t; t++)
count += (*t == ch);
size_t rlen = strlen(repl);
char *res = (char*) mCalloc ( (int) strlen(s) + ( ( (int)rlen - 1) * count) + 1, sizeof(char));
char *ptr = res;
for(t=s; *t; t++)
{
if(*t == ch)
{
memcpy(ptr, repl, rlen);
ptr += rlen;
}
else
{
*ptr++ = *t;
}
}
*ptr = 0;
return res;
}
int ludcmp(double **a, int n, double *d)
{
int i,imax,j,k;
double big,dum,sum,temp;
double *vv;
imax = 0;
vv = (double *)mCalloc(n,sizeof(double));
*d=1.0;
for (i=0;i<n;i++)
{
big=0.0;
for (j=0;j<n;j++)
if ((temp=fabs(a[i][j])) > big) big=temp;
if (big == 0.0) Exit("\n. Singular matrix in routine LUDCMP");
vv[i]=1.0/big;
}
for (j=0;j<n;j++)
{
for (i=0;i<j;i++)
{
sum=a[i][j];
for (k=0;k<i;k++) sum -= a[i][k]*a[k][j];
a[i][j]=sum;
}
big=0.0;
for (i=j;i<n;i++) {
sum=a[i][j];
for (k=0;k<j;k++)
sum -= a[i][k]*a[k][j];
a[i][j]=sum;
if ((dum=vv[i]*fabs(sum)) >= big)
{
big=dum;
imax=i;
}
}
if (j != imax)
{
for (k=0;k<n;k++)
{
dum=a[imax][k];
a[imax][k]=a[j][k];
a[j][k]=dum;
}
*d = -(*d);
vv[imax]=vv[j];
}
if (a[j][j] == 0.0) a[j][j]=LUDCMP_TINY;
if (j != n) {
dum=1.0/(a[j][j]);
for (i=j+1;i<n;i++) a[i][j] *= dum;
}
}
Free(vv);
return(0);
}
int *initOneArray (int l)
{
int *x;
int i;
x = (int *) mCalloc (l, sizeof (int));
for (i=0; i<l; i++)
x[i] = 1;
return (x);
}
char *getLine (FILE *file, char *line, const int len)
{
line = (char*) mCalloc (len, sizeof(char));
memset (line, '\0', (size_t) len);
if (! fgets (line, len, file))
{
Exit ( (char*)"Cannot read line.");
}
return line;
}
void Usage()
{
char *BOLD=(char *)mCalloc(10,sizeof(char));
char *FLAT=(char *)mCalloc(10,sizeof(char));
char *LINE=(char *)mCalloc(10,sizeof(char));
char *cha;
cha =getenv("OS");
if(cha!=NULL)
{
strcpy(BOLD, "");
strcpy(FLAT, "");
strcpy(LINE, "");
}
else
{
strcpy(BOLD, "\033[00;01m");
strcpy(FLAT, "\033[00;00m");
strcpy(LINE, "\033[00;04m");
}
#ifdef PHYML
PhyML_Printf("%sNAME\n"
"%s\t- PhyML %s - \n\n"
"%s\t\''A simple, fast, and accurate algorithm to estimate\n"
"%s\tlarge phylogenies by maximum likelihood\''\n\n"
"%s\tStephane Guindon and Olivier Gascuel,\n"
"%s\tSystematic Biology 52(5):696-704, 2003.\n\n"
"%s\tPlease cite this paper if you use this software in your publications.\n",BOLD,FLAT,VERSION,FLAT,FLAT,FLAT,FLAT,FLAT);
#endif
#ifdef PHYTIME
PhyML_Printf("%sNAME\n"
"%s\t- PhyTime %s - \n\n"
"%s\t'Bayesian estimation of divergence times from large sequence alignments.'\n"
"%s\tStephane Guindon,\n"
"%s\tMolecular Biology and Evolution 27(8):1768-81, 2010.\n\n"
"%s\tPlease cite this paper if you use this software in your publications.\n",BOLD,FLAT,VERSION,FLAT,FLAT,FLAT,FLAT,FLAT);
#endif
#ifdef PHYML
PhyML_Printf("%s\nSYNOPSIS:\n\n"
"%s\tphyml %s[command args]\n",BOLD,BOLD,BOLD);
#endif
#ifdef PHYTIME
PhyML_Printf("%s\nSYNOPSIS:\n\n"
"%s\tphytime %s[command args]\n",BOLD,BOLD,BOLD);
#endif
#ifdef PHYML
PhyML_Printf("%s\n\tAll the options below are optional (except '%s-i%s' if you want to use the command-line interface).\n\n",FLAT,BOLD,FLAT);
#endif
#ifdef PHYTIME
PhyML_Printf("%s\n\tAll the options below are optional except '%s-i%s','%s-u%s' and '%s--calibration%s'.\n\n",FLAT,BOLD,FLAT,BOLD,FLAT,BOLD,FLAT);
#endif
PhyML_Printf("%s\nCommand options:\n%s",BOLD,FLAT);
PhyML_Printf("\n\t%s-i (or --input) %sseq_file_name%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\t%sseq_file_name%s is the name of the nucleotide or amino-acid sequence file in PHYLIP format.\n",LINE,FLAT);
PhyML_Printf("\n");
PhyML_Printf("%s\n\t-d (or --datatype) ""%sdata_type%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\t%sdata_type%s is 'nt' for nucleotide (default), 'aa' for amino-acid sequences, or 'generic',\n",LINE,FLAT);
PhyML_Printf("\t\t(use NEXUS file format and the 'symbols' parameter here).\n");
PhyML_Printf("\n");
PhyML_Printf("%s\n\t-q (or --sequential)%s\n",BOLD,FLAT);
PhyML_Printf("%s\t\tChanges interleaved format (default) to sequential format.\n",FLAT);
PhyML_Printf("\n");
#ifndef PHYTIME
PhyML_Printf("%s\n\t-n (or --multiple) ""%snb_data_sets%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\t%snb_data_sets%s is an integer corresponding to the number of data sets to analyse.\n",LINE,FLAT);
PhyML_Printf("\n");
#endif
#ifndef PHYTIME
PhyML_Printf("%s\n\t-p (or --pars)%s\n",BOLD,FLAT);
PhyML_Printf("%s\t\tUse a minimum parsimony starting tree. This option is taken into account when the '-u' option\n",FLAT);
PhyML_Printf("%s\t\tis absent and when tree topology modifications are to be done.\n",FLAT);
PhyML_Printf("\n");
#endif
#ifndef PHYTIME
PhyML_Printf("%s\n\t-b (or --bootstrap) %sint%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\t%sint%s > 0: %sint%s is the number of bootstrap replicates.\n",LINE,FLAT,LINE,FLAT);
PhyML_Printf("\t\t%sint%s = 0: neither approximate likelihood ratio test nor bootstrap values are computed.\n",LINE,FLAT);
PhyML_Printf("\t\t%sint%s = -1: approximate likelihood ratio test returning aLRT statistics.\n",LINE,FLAT);
PhyML_Printf("\t\t%sint%s = -2: approximate likelihood ratio test returning Chi2-based parametric branch supports.\n",LINE,FLAT);
/* PhyML_Printf("\t\t%sint%s = -3 : minimum of Chi2-based parametric and SH-like branch supports.\n",LINE,FLAT); */
PhyML_Printf("\t\t%sint%s = -4: (default) SH-like branch supports alone.\n",LINE,FLAT);
PhyML_Printf("\t\t%sint%s = -5: approximate Bayes branch supports.\n",LINE,FLAT);
PhyML_Printf("\n");
#endif
PhyML_Printf("%s\n\t-m (or --model) %smodel%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tmodel%s : substitution model name.\n",FLAT);
PhyML_Printf("\t\t%s- %sNucleotide%s-based models : %sHKY85%s (default) | %sJC69%s | %sK80%s | %sF81%s | %sF84%s | %sTN93%s | %sGTR%s | %scustom (*)%s\n",
FLAT,LINE,FLAT,LINE,FLAT,LINE,FLAT,LINE,FLAT,LINE,FLAT,LINE,FLAT,LINE,FLAT,LINE,FLAT,LINE,FLAT);
PhyML_Printf("\t\t(*) : for the custom option, a string of six digits identifies the model. For instance, 000000\n");
PhyML_Printf("\t\t corresponds to F81 (or JC69 provided the distribution of nucleotide frequencies is uniform).\n");
PhyML_Printf("\t\t 012345 corresponds to GTR. This option can be used for encoding any model that is a nested within GTR.\n");
PhyML_Printf("\n");
PhyML_Printf("\t\t%s- %sAmino-acid%s based models : %sLG%s (default) | %sWAG%s | %sJTT%s | %sMtREV%s | %sDayhoff%s | %sDCMut%s | %sRtREV%s | %sCpREV%s | %sVT%s\n",
FLAT,LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT);
PhyML_Printf("\t\t %sBlosum62%s | %sMtMam%s | %sMtArt%s | %sHIVw%s | %sHIVb%s | %scustom%s\n",
LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT,
LINE,FLAT);
PhyML_Printf("\n");
PhyML_Printf("%s\n\t--aa_rate_file %sfilename%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\t%sfilename%s is the name of the file that provides the amino acid substitution rate matrix in PAML format.\n",LINE,FLAT);
PhyML_Printf("\t\tIt is compulsory to use this option when analysing amino acid sequences with the `custom' model.\n");
PhyML_Printf("\n");
#ifndef PHYTIME
PhyML_Printf("%s\n\t-f %se%s, %sm%s, or %sfA,fC,fG,fT%s\n",BOLD,LINE,BOLD,LINE,BOLD,LINE,FLAT);
PhyML_Printf("\t\t%se%s : the character frequencies are determined as follows : \n",LINE,FLAT);
PhyML_Printf("%s\t\t- %sNucleotide%s sequences: (Empirical) the equilibrium base frequencies are estimated by counting\n"
"\t\t the occurence of the different bases in the alignment.\n",FLAT,LINE,FLAT);
PhyML_Printf("%s\t\t- %sAmino-acid%s sequences: (Empirical) the equilibrium amino-acid frequencies are estimated by counting\n"
"\t\t the occurence of the different amino-acids in the alignment.\n",FLAT,LINE,FLAT);
PhyML_Printf("\n");
PhyML_Printf("\t\t%sm%s : the character frequencies are determined as follows : \n",LINE,FLAT);
PhyML_Printf("%s\t\t- %sNucleotide%s sequences: (ML) the equilibrium base frequencies are estimated using maximum likelihood \n",FLAT,LINE,FLAT);
PhyML_Printf("%s\t\t- %sAmino-acid%s sequences: (Model) the equilibrium amino-acid frequencies are estimated using\n"
"\t\t the frequencies defined by the substitution model.\n",FLAT,LINE,FLAT);
PhyML_Printf("\n");
PhyML_Printf("\t\t%s\"fA,fC,fG,fT\"%s : only valid for nucleotide-based models. fA, fC, fG and fT are floating numbers that \n",LINE,FLAT);
PhyML_Printf("\t\t correspond to the frequencies of A, C, G and T respectively (WARNING: do not use any blank space between\n");
PhyML_Printf("\t\t your values of nucleotide frequencies, only commas!)\n");
PhyML_Printf("\n");
#endif
#ifdef PHYTIME
PhyML_Printf("%s\n\t--calibration %sfilename%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\t%sfilename%s is the name of the calibration file that provides a priori defined boundaries for node ages.\n",LINE,FLAT);
PhyML_Printf("\t\tPlease read the manual for more information about the format of this file.\n");
PhyML_Printf("\n");
#endif
PhyML_Printf("%s\n\t-t (or --ts/tv) %sts/tv_ratio%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tts/tv_ratio%s : transition/transversion ratio. DNA sequences only.\n",FLAT);
PhyML_Printf("\t\tCan be a fixed positive value (ex:4.0) or %se%s to get the maximum likelihood estimate.\n",LINE,FLAT);
PhyML_Printf("\n");
PhyML_Printf("%s\n\t-v (or --pinv) %sprop_invar%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tprop_invar%s : proportion of invariable sites.\n",FLAT);
PhyML_Printf("\t\tCan be a fixed value in the [0,1] range or %se%s to get the maximum likelihood estimate.\n",LINE,FLAT);
PhyML_Printf("\n");
PhyML_Printf("%s\n\t-c (or --nclasses) %snb_subst_cat%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tnb_subst_cat%s : number of relative substitution rate categories. Default : %snb_subst_cat%s=4.\n",
FLAT,LINE,FLAT);
PhyML_Printf("\t\tMust be a positive integer.\n");
PhyML_Printf("\n");
PhyML_Printf("%s\n\t-a (or --alpha) %sgamma%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tgamma%s : distribution of the gamma distribution shape parameter.\n",FLAT);
PhyML_Printf("\t\tCan be a fixed positive value or %se%s to get the maximum likelihood estimate.\n",LINE,FLAT);
PhyML_Printf("\n");
#ifndef PHYTIME
PhyML_Printf("%s\n\t-s (or --search) %smove%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tTree topology search operation option.\n");
PhyML_Printf("\t\tCan be either %sNNI%s (default, fast) or %sSPR%s (a bit slower than NNI) or %sBEST%s (best of NNI and SPR search).\n",LINE,FLAT,LINE,FLAT,LINE,FLAT);
PhyML_Printf("\n");
#endif
PhyML_Printf("%s\n\t-u (or --inputtree) %suser_tree_file%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tuser_tree_file%s : starting tree filename. The tree must be in Newick format.\n",FLAT);
PhyML_Printf("\n");
#ifndef PHYTIME
PhyML_Printf("%s\n\t-o %sparams%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tThis option focuses on specific parameter optimisation.\n");
PhyML_Printf("\t\t%sparams%s=tlr : tree topology (t), branch length (l) and rate parameters (r) are optimised.\n",LINE,FLAT);
PhyML_Printf("\t\t%sparams%s=tl : tree topology and branch length are optimised.\n",LINE,FLAT);
PhyML_Printf("\t\t%sparams%s=lr : branch length and rate parameters are optimised.\n",LINE,FLAT);
PhyML_Printf("\t\t%sparams%s=l : branch length are optimised.\n",LINE,FLAT);
PhyML_Printf("\t\t%sparams%s=r : rate parameters are optimised.\n",LINE,FLAT);
PhyML_Printf("\t\t%sparams%s=n : no parameter is optimised.\n",LINE,FLAT);
PhyML_Printf("\n");
#endif
#ifndef PHYTIME
PhyML_Printf("%s\n\t--rand_start%s\n",BOLD,FLAT);
PhyML_Printf("\t\tThis option sets the initial tree to random.\n");
PhyML_Printf("\t\tIt is only valid if SPR searches are to be performed.\n");
PhyML_Printf("\n");
#endif
#ifndef PHYTIME
PhyML_Printf("%s\n\t--n_rand_starts %snum%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tnum%s is the number of initial random trees to be used.\n",FLAT);
PhyML_Printf("\t\tIt is only valid if SPR searches are to be performed.\n");
PhyML_Printf("\n");
#endif
PhyML_Printf("%s\n\t--r_seed %snum%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tnum%s is the seed used to initiate the random number generator.\n",FLAT);
PhyML_Printf("\t\tMust be an integer.\n");
PhyML_Printf("\n");
#ifndef PHYTIME
PhyML_Printf("%s\n\t--print_site_lnl%s\n",BOLD,FLAT);
PhyML_Printf("\t\t%sPrint the likelihood for each site in file *_phyml_lk.txt.\n",FLAT);
PhyML_Printf("\n");
#endif
#ifndef PHYTIME
PhyML_Printf("%s\n\t--print_trace%s\n",BOLD,FLAT);
PhyML_Printf("\t\t%sPrint each phylogeny explored during the tree search process\n",FLAT);
PhyML_Printf("\t\t%sin file *_phyml_trace.txt.\n",FLAT);
PhyML_Printf("\n");
#endif
PhyML_Printf("%s\n\t--run_id %sID_string%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\t%sAppend the string %sID_string%s at the end of each PhyML output file.\n",FLAT,LINE,FLAT);
PhyML_Printf("\t\t%sThis option may be useful when running simulations involving PhyML.\n",FLAT);
PhyML_Printf("\n");
PhyML_Printf("%s\n\t--quiet%s\n",BOLD,FLAT);
PhyML_Printf("\t\t%sNo interactive question (for running in batch mode) and quiet output.\n",FLAT);
PhyML_Printf("\n");
PhyML_Printf("%s\n\t--no_memory_check%s\n",BOLD,FLAT);
PhyML_Printf("\t\t%sNo interactive question for memory usage (for running in batch mode). Normal ouput otherwise.\n",FLAT);
PhyML_Printf("\n");
#ifndef PHYTIME
PhyML_Printf("%s\n\t--alias_subpatt%s\n",BOLD,FLAT);
PhyML_Printf("\t\t%sSite aliasing is generalized at the subtree level. Sometimes lead to faster calculations.\n",FLAT);
PhyML_Printf("\t\t%sSee Kosakovsky Pond SL, Muse SV, Sytematic Biology (2004) for an example.\n",FLAT);
PhyML_Printf("\n");
#endif
#ifndef PHYTIME
PhyML_Printf("%s\n\t--boot_progress_display %snum%s (default=20)\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\t%snum%s is the frequency at which the bootstrap progress bar will be updated.\n",LINE,FLAT);
PhyML_Printf("\t\tMust be an integer.\n");
PhyML_Printf("\n");
#endif
#ifdef PHYTIME
PhyML_Printf("%s\n\t--chain_len %snum%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\t%snum%s is the number of generations or runs of the Markov Chain Monte Carlo. Set to 1E+6 by default. \n",LINE,FLAT);
PhyML_Printf("\t\tMust be an integer.\n");
PhyML_Printf("\n");
#endif
/* #ifdef PHYTIME */
/* PhyML_Printf("%s\n\t--burnin %snum%s\n",BOLD,LINE,FLAT); */
/* PhyML_Printf("\t\t%snum%s is the number of generations of runs of the Markov Chain Monte Carlo during the 'burnin' period.\n",LINE,FLAT); */
/* PhyML_Printf("\t\t%sSet to 1E+5 by default. Must be an integer. \n",FLAT); */
/* PhyML_Printf("\n"); */
/* #endif */
#ifdef PHYTIME
PhyML_Printf("%s\n\t--sample_freq %snum%s\n",BOLD,LINE,FLAT);
PhyML_Printf("\t\tThe chain is sampled every %snum%s generations. Set to 1E+3 by default. \n",LINE,FLAT);
PhyML_Printf("\t\tMust be an integer.\n");
PhyML_Printf("\n");
#endif
#ifdef PHYTIME
PhyML_Printf("%s\n\t--no_data%s\n",BOLD,FLAT);
PhyML_Printf("\t\tUse this option to sample from the priors only (rather from the posterior joint density\n");
PhyML_Printf("\t\tof the model parameters).\n");
PhyML_Printf("\n");
#endif
#ifdef PHYTIME
PhyML_Printf("%s\n\t--fastlk%s\n",BOLD,FLAT);
PhyML_Printf("\t\tUse the multivariate normal approximation to the likelihood and speed up calculations\n");
PhyML_Printf("\n");
#endif
#ifdef PHYML
PhyML_Printf("%sPHYLIP-LIKE INTERFACE\n""%s\n\tYou can also use PhyML with no argument, in this case change the value of\n",BOLD,FLAT);
PhyML_Printf("%s\ta parameter by typing its corresponding character as shown on screen.\n\n",FLAT);
#endif
#ifdef PHYML
PhyML_Printf("%sEXAMPLES\n\n"
"%s\tDNA interleaved sequence file, default parameters : ""%s ./phyml -i seqs1"
"%s\n\tAA interleaved sequence file, default parameters : ""%s ./phyml -i seqs2 -d aa"
"%s\n\tAA sequential sequence file, with customization : ""%s ./phyml -i seqs3 -q -d aa -m JTT -c 4 -a e%s\n",BOLD,FLAT,BOLD,FLAT,BOLD,FLAT,BOLD,FLAT);
#endif
Exit("");
}
void SPR (tree *T, double **D, double **A, int *count, FILE *statfile)
{
int i, j, k;
node *v;
edge *e, *f;
double ***swapWeights;
double treeWeightBefore;
double swapValue = 0.0;
boolean firstSPR = TRUE;
swapWeights = (double ***) mCalloc (2, sizeof(double **));
makeBMEAveragesTable (T, D, A);
assignBMEWeights (T, A);
weighTree (T);
treeWeightBefore = T->weight;
for (i=0; i<2; i++)
swapWeights[i] = initDoubleMatrix (T->size);
do
{
swapValue = 0.0;
zero3DMatrix (swapWeights, 2, T->size, T->size);
i = j = k = 0;
for (e = depthFirstTraverse (T, NULL); NULL != e; e = depthFirstTraverse (T, e))
assignSPRWeights (e->head, A, swapWeights);
findTableMin (&i, &j, &k, T->size, swapWeights, &swapValue);
swapValue = swapWeights[i][j][k];
if (swapValue < -FLT_EPSILON)
{
if (firstSPR)
{
firstSPR = FALSE;
if (!isBoostrap)
{
if (verbose > 2)
Debug ( (char*)"Before SPR: tree length is %f.", treeWeightBefore);
else if (verbose > 1)
Message ( (char*)". Before SPR: tree length is %f.", treeWeightBefore);
}
}
if (verbose > 2 && !isBoostrap)
Debug ( (char*)"New tree length should be %f.", T->weight + 0.25 * swapValue);
v = indexedNode (T, j);
f = indexedEdge (T, k);
if (verbose > 2 && !isBoostrap)
{
if ((NULL == f->head->label) || (strlen (f->head->label) == 0))
{
if ((NULL == f->tail->label) || (strlen (f->tail->label) == 0))
Debug ( (char*)"Swapping tree below '%s' to split edge '%s' with internal head and tail.",
v->parentEdge->label, f->label);
else
Debug ( (char*)"Swapping tree below '%s' to split edge '%s' with internal head and tail '%s'.",
v->parentEdge->label, f->label, f->tail->label);
}
else
{
if ((NULL == f->tail->label) || (strlen (f->tail->label) == 0))
Debug ( (char*)"Swapping tree below '%s' to split edge '%s' with head '%s' and internal tail.",
v->parentEdge->label, f->label, f->head->label, f->tail->label);
else
Debug ( (char*)"Swapping tree below '%s' to split edge '%s' with head '%s' and tail '%s'.",
v->parentEdge->label, f->label, f->head->label, f->tail->label);
}
}
SPRTopShift (v, f, 2-i);
makeBMEAveragesTable (T, D, A);
assignBMEWeights (T, A);
weighTree (T);
(*count)++;
if (!isBoostrap)
{
if (verbose > 2)
Debug ( (char*)"SPR %5d: new tree length is %f.", *count, T->weight);
else if (verbose > 1)
Message ( (char*)". SPR %5d: new tree length is %f.", *count, T->weight);
}
}
} while (swapValue < -FLT_EPSILON);
for (i=0; i<2; i++)
freeMatrix (swapWeights[i], T->size);
free (swapWeights);
return;
}
/* An inelegant iterative version */
void SPRUpShift (node *vmove, edge *esplit)
{
edge *f;
edge **EPath;
edge **sib;
node **v;
int i;
int pathLength;
for (f=esplit->tail->parentEdge,pathLength=1; f->tail != vmove; f=f->tail->parentEdge)
pathLength++;
/* count number of edges to vmove
* note that pathLength > 0 will hold */
EPath = (edge **) mCalloc (pathLength, sizeof(edge *));
v = (node **) mCalloc (pathLength, sizeof(edge *));
sib = (edge **) mCalloc ((pathLength+1), sizeof(edge *));
/* there are pathLength + 1 side trees, one at the head and tail of each edge in the path */
sib[pathLength] = siblingEdge (esplit);
i = pathLength;
f = esplit->tail->parentEdge;
while (i > 0)
{
i--;
EPath[i] = f;
sib[i] = siblingEdge (f);
v[i] = f->head;
f = f->tail->parentEdge;
}
/* indexed so head of Epath is v[i], tail is v[i-1] and sibling edge is sib[i]
* need to assign head, tail of each edge in path
* as well as have proper values for the left and right fields */
if (esplit == esplit->tail->leftEdge)
{
vmove->leftEdge = esplit;
vmove->rightEdge = EPath[pathLength-1];
}
else
{
vmove->rightEdge = esplit;
vmove->leftEdge = EPath[pathLength-1];
}
esplit->tail = vmove;
/* espilt->head remains unchanged
* vmove has the proper fields for left, right, and parentEdge */
for(i=0; i<(pathLength-1); i++)
EPath[i]->tail = v[i+1];
/* this bit flips the orientation along the path
* the tail of Epath[i] is now v[i+1] instead of v[i-1] */
EPath[pathLength-1]->tail = vmove;
for (i=1; i<pathLength; i++)
{
if (sib[i+1] == v[i]->leftEdge)
v[i]->rightEdge = EPath[i-1];
else
v[i]->leftEdge = EPath[i-1];
}
if (sib[1] == v[0]->leftEdge)
v[0]->rightEdge = sib[0];
else
v[0]->leftEdge = sib[0];
sib[0]->tail = v[0];
free(EPath);
free(v);
free(sib);
return;
}
/**
* Fill the Option fields, with the argc array
*/
int Read_Command_Line(option *io, int argc, char **argv)
{
int c;
int idx;
int i;
int writemode;
PhyML_Printf("\n. command-line: ");
For(i,argc) PhyML_Printf("%s ",argv[i]);
if(argc == 1) Exit("\n. No argument was passed to the program. Please check the documentation. \n");
struct option longopts[] =
{
{"n_rgrft", required_argument,NULL,0},
{"n_globl", required_argument,NULL,1},
{"max_dist", required_argument,NULL,2},
{"n_optim", required_argument,NULL,3},
{"n_best", required_argument,NULL,4},
{"model", required_argument,NULL,5},
{"search", required_argument,NULL,6},
{"datatype", required_argument,NULL,7},
{"multiple", required_argument,NULL,8},
{"input", required_argument,NULL,9},
{"bootstrap", required_argument,NULL,10},
{"ts/tv", required_argument,NULL,11},
{"nclasses", required_argument,NULL,12},
{"pinv", required_argument,NULL,13},
{"alpha", required_argument,NULL,14},
{"inputtree", required_argument,NULL,15},
{"min_diff_lk_local", required_argument,NULL,16},
{"min_diff_lk_global",required_argument,NULL,17},
{"steph_spr", no_argument,NULL,18},
{"brent_it_max", required_argument,NULL,19},
{"rand_start", no_argument,NULL,20},
{"n_rand_starts", required_argument,NULL,21},
{"sequential", no_argument,NULL,22},
{"inside_opt", no_argument,NULL,23},
{"p_moves", required_argument,NULL,24},
{"fast_nni", no_argument,NULL,25},
{"g_pars", no_argument,NULL,26},
{"r_seed", required_argument,NULL,27},
{"collapse_boot", required_argument,NULL,28},
{"random_boot", required_argument,NULL,29},
{"print_trace", no_argument,NULL,30},
{"print_site_lnl", no_argument,NULL,31},
{"print_site_lk", no_argument,NULL,31},
{"cov", no_argument,NULL,32},
{"cov_delta", required_argument,NULL,33},
{"cov_alpha", required_argument,NULL,34},
{"cov_ncats", required_argument,NULL,35},
{"ps", no_argument,NULL,36},
{"cov_free", no_argument,NULL,37},
{"no_gap", no_argument,NULL,38},
{"n_rr_branch", required_argument,NULL,39},
{"append", no_argument,NULL,40},
{"no_five_branch", no_argument,NULL,41},
{"pars_thresh", required_argument,NULL,42},
{"min_diff_lk_move", required_argument,NULL,43},
{"hybrid", no_argument,NULL,44},
{"use_median", no_argument,NULL,45},
{"run_id", required_argument,NULL,46},
{"pars", no_argument,NULL,47},
{"quiet", no_argument,NULL,48},
{"version", no_argument,NULL,49},
{"calibration_file", required_argument,NULL,50},
{"calibration", required_argument,NULL,50},
{"clade_file", required_argument,NULL,50},
{"boot_progress_every", required_argument,NULL,51},
{"aa_rate_file", required_argument,NULL,52},
{"chain_len", required_argument,NULL,53},
{"sample_freq", required_argument,NULL,54},
{"burnin", required_argument,NULL,55},
{"no_memory_check", no_argument,NULL,56},
{"no_colalias", no_argument,NULL,57},
{"alias_subpatt", no_argument,NULL,58},
{"no_sequences", no_argument,NULL,59},
{"prior", no_argument,NULL,59},
{"fastlk", no_argument,NULL,60},
{"free_rates", no_argument,NULL,61},
{"freerates", no_argument,NULL,61},
{"freerate", no_argument,NULL,61},
{"free_rate", no_argument,NULL,61},
{"is", no_argument,NULL,62},
// no 63 since it corresponds to character '?'
{"rate_model", required_argument,NULL,64},
{"ratemodel", required_argument,NULL,64},
{"log_l", no_argument,NULL,65},
{"gamma_lens", no_argument,NULL,66},
{"il", no_argument,NULL,66},
{"codpos", required_argument,NULL,67},
{"constraint_file", required_argument,NULL,68},
{"constraint_tree", required_argument,NULL,68},
{"help", no_argument,NULL,69},
{"mutmap", no_argument,NULL,70},
{"parvals", required_argument,NULL,71},
{"constrained_lens", no_argument,NULL,72},
{"xml", required_argument,NULL,73},
{"l_var", required_argument,NULL,74},
#ifdef BEAGLE
{"beagle_resource", required_argument,NULL,75},
#endif
{"ancestral", no_argument,NULL,76},
{"anc", no_argument,NULL,76},
{"coord_file", required_argument,NULL,77},
{0,0,0,0}
};
io->datatype = UNDEFINED;
#ifndef PHYML
int open_ps_file = 0;
#endif
idx=-1;
do
{
c = getopt_long(argc,argv,"qi:d:m:b:n:t:f:zk:v:c:a:u:ho:s:x:g:l:ep",longopts,&idx);
switch(c)
{
case 77:
{
char *tmp;
tmp = (char *)mCalloc(T_MAX_FILE, sizeof(char));
if(strlen(optarg) > T_MAX_FILE -11)
{
char choix;
strcpy (tmp, "\n. The file name'");
strcat (tmp, optarg);
strcat (tmp, "' is too long.\n");
PhyML_Printf("%s",tmp);
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else if (!Filexists (optarg))
{
char choix;
strcpy (tmp, "\n. The file '");
strcat (tmp, optarg);
strcat (tmp, "' doesn't exist.\n");
PhyML_Printf("%s",tmp);
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else
{
strcpy(io->in_coord_file, optarg);
io->fp_in_coord = Openfile(io->in_coord_file,READ);
}
Free(tmp);
break;
}
case 76:
{
io->ancestral = YES;
break;
}
#ifdef BEAGLE
case 75:
{
io->beagle_resource = (int)atoi(optarg);
break;
}
#endif
case 74:
{
io->mod->l_var_sigma = String_To_Dbl(optarg);
break;
}
case 73:
{
#ifdef INVITEE
Free_Optimiz(io->mod->s_opt);
M4_Free_M4_Model(io->mod->m4mod);
Free_Model_Basic(io->mod);
Free_Input(io);
PhyTime_XML(optarg);
return 0;
#elif defined(PHYML)
Free_Optimiz(io->mod->s_opt);
M4_Free_M4_Model(io->mod->m4mod);
Free_Model_Basic(io->mod);
Free_Input(io);
PhyML_XML(optarg);
return 0;
#elif defined(DATE)
Free_Optimiz(io->mod->s_opt);
M4_Free_M4_Model(io->mod->m4mod);
Free_Model_Basic(io->mod);
Free_Input(io);
DATE_XML(optarg);
return 0;
#endif
break;
}
case 72:
{
io->mod->s_opt->constrained_br_len = YES;
break;
}
case 71:
{
io->mcmc->in_fp_par = fopen(optarg,"r");
io->mcmc->randomize = NO;
break;
}
case 70:
{
io->mutmap = YES;
break;
}
case 68:
{
char *tmp;
tmp = (char *)mCalloc(T_MAX_FILE, sizeof(char));
if(strlen(optarg) > T_MAX_FILE -11)
{
char choix;
strcpy (tmp, "\n. The file name'");
strcat (tmp, optarg);
strcat (tmp, "' is too long.\n");
PhyML_Printf("%s",tmp);
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else if (!Filexists (optarg))
{
char choix;
strcpy (tmp, "\n. The file '");
strcat (tmp, optarg);
strcat (tmp, "' doesn't exist.\n");
PhyML_Printf("%s",tmp);
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else
{
strcpy(io->in_constraint_tree_file, optarg);
io->fp_in_constraint_tree = Openfile(io->in_constraint_tree_file,0);
}
Free(tmp);
break;
}
case 67:
{
phydbl pos;
pos = atof(optarg);
io->codpos = (int)pos;
if(io->codpos < 1 || io->codpos > 3)
{
char choix;
PhyML_Printf("\n. Coding position must be set to 1, 2 or 3.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
break;
break;
}
case 66:
{
io->mod->gamma_mgf_bl = YES;
io->mod->s_opt->opt_gamma_br_len = YES;
break;
}
case 65:
{
io->mod->log_l = YES;
break;
}
case 64:
{
char *s;
int i;
s = (char *)mCalloc(T_MAX_NAME,sizeof(char));
i = 0;
while(optarg[i++]) s[i]=tolower(optarg[i]);
if(!strcmp(optarg,"gbd")) io->rates->model = THORNE;
else if(!strcmp(optarg,"gbs")) io->rates->model = GUINDON;
else if(!strcmp(optarg,"gamma")) io->rates->model = GAMMA;
else if(!strcmp(optarg,"clock")) io->rates->model = STRICTCLOCK;
else if(!strcmp(optarg,"strictclock")) io->rates->model = STRICTCLOCK;
else if(!strcmp(optarg,"strict_clock")) io->rates->model = STRICTCLOCK;
else
{
PhyML_Printf("\n. rate_model should be 'gbs', 'gbd', 'gamma' or 'clock'.");
Exit("\n");
}
Free(s);
break;
}
case 62:
{
io->mcmc->is = YES;
break;
}
case 61:
{
io->mod->ras->free_mixt_rates = YES;
io->mod->s_opt->opt_free_mixt_rates = YES;
break;
}
case 60:
{
io->lk_approx = NORMAL;
break;
}
case 59:
{
io->mcmc->use_data = NO;
break;
}
case 58:
{
io->do_alias_subpatt = YES;
break;
}
case 57:
{
io->colalias = NO;
break;
}
case 56:
{
io->mem_question = NO;
break;
}
case 55:
{
phydbl len;
len = atof(optarg);
io->mcmc->chain_len_burnin = (int)len;
if(io->mcmc->chain_len_burnin < 1)
{
char choix;
PhyML_Printf("\n. chain_len_burnin must be an integer greater than 0.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
break;
}
case 54:
{
phydbl len;
len = atof(optarg);
io->mcmc->sample_interval = (int)len;
if(io->mcmc->sample_interval < 1)
{
char choix;
PhyML_Printf("\n. sample_interval must be an integer greater than 0.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
break;
}
case 53:
{
phydbl len;
len = atof(optarg);
io->mcmc->chain_len = (int)len;
if(io->mcmc->chain_len < 1)
{
char choix;
PhyML_Printf("\n. chain_len must be an integer greater than 0.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
break;
}
case 52:
{
char *s;
s = (char *)mCalloc(T_MAX_FILE, sizeof(char));
strcpy(s,optarg);
io->mod->fp_aa_rate_mat = Openfile(s,0);
strcpy(io->mod->aa_rate_mat_file->s,s);
Free(s);
break;
}
case 51:
{
io->boot_prog_every = atoi(optarg);
if(io->boot_prog_every < 1)
{
char choix;
PhyML_Printf("\n. boot_progress_every must be an integer greater than 0.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
break;
}
case 50:
{
strcpy(io->clade_list_file,optarg);
break;
}
case 49:
{
PhyML_Printf("\n. This is PhyML version %s.\n\n",VERSION);
Exit("");
break;
}
case 48 :
{
io->quiet = 1;
break;
}
case 'p' : case 47 :
{
io->in_tree = 1;
break;
}
case 46 :
{
io->append_run_ID = YES;
strcpy(io->run_id_string,optarg);
break;
}
case 45 :
{
io->mod->ras->gamma_median = 1;
break;
}
case 44 :
{
io->mod->s_opt->hybrid_thresh = 0;
break;
}
case 43 :
{
io->mod->s_opt->min_diff_lk_move = atof(optarg);
if(io->mod->s_opt->min_diff_lk_move < 0)
{
char choix;
PhyML_Printf("\n. Min_diff_lk_move must be a double greater than 0.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
break;
}
case 42 :
{
io->mod->s_opt->pars_thresh = (int)atoi(optarg);
if(io->mod->s_opt->pars_thresh < 0)
{
PhyML_Printf("\n. The parsimony threshold must be an integer greater than 0.\n");
PhyML_Printf("\n. Type any key to exit.\n");
Exit("\n");
}
break;
}
case 41 :
{
io->mod->s_opt->opt_five_branch = 0;
break;
}
case 40 :
{
writemode = 2;
break;
}
case 39 :
{
break;
}
case 38 :
{
io->rm_ambigu = 1;
break;
}
case 37 :
{
io->mod->s_opt->opt_cov_free_rates = YES;
io->mod->m4mod->use_cov_alpha = NO;
io->mod->m4mod->use_cov_free = YES;
break;
}
case 36 :
{
#ifndef PHYML
open_ps_file = 1;
#endif
break;
}
case 35 :
{
io->mod->m4mod->n_h = (int)atoi(optarg);
if(io->mod->m4mod->n_h < 1)
{
char choix;
PhyML_Printf("\n. The number of classes must be greater than 0.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
break;
}
case 34 :
{
io->mod->m4mod->use_cov_alpha = YES;
io->mod->m4mod->use_cov_free = NO;
if(!strcmp(optarg,"e") || !strcmp(optarg,"E") ||
!strcmp(optarg,"estimated") || !strcmp(optarg,"ESTIMATED"))
{
io->mod->s_opt->opt_cov_alpha = YES;
io->mod->m4mod->alpha = 1.0;
}
else
{
io->mod->m4mod->alpha = (phydbl)atof(optarg);
if(io->mod->m4mod->alpha < 1.E-5)
{
char choix;
PhyML_Printf("\n. The value of alpha must be greater than 1.E-5.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
}
break;
}
case 33 :
{
if(!strcmp(optarg,"e") || !strcmp(optarg,"E") ||
!strcmp(optarg,"estimated") || !strcmp(optarg,"ESTIMATED"))
{
io->mod->s_opt->opt_cov_delta = YES;
io->mod->m4mod->delta = 1.0;
}
else
{
io->mod->m4mod->delta = (phydbl)atof(optarg);
if(atof(optarg) < 1.E-10)
{
char choix;
PhyML_Printf("\n. The value of delta must be larger than 1.E-10.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
}
break;
}
case 32 :
{
io->mod->use_m4mod = YES;
break;
}
case 31 :
{
io->print_site_lnl = YES;
break;
}
case 30 :
{
io->print_trace = YES;
break;
}
case 29 :
{
io->random_boot_seq_order = (int)atoi(optarg);
break;
}
case 28 :
{
io->collapse_boot = (int)atoi(optarg);
break;
}
case 27 :
{
io->r_seed = (int)atoi(optarg);
break;
}
case 26 :
{
io->mod->s_opt->general_pars = YES;
break;
}
case 25 :
{
io->mod->s_opt->fast_nni = YES;
break;
}
case 24 :
{
io->mod->s_opt->p_moves_to_examine = (phydbl)atof(optarg);
break;
}
case 23 :
{
io->mod->s_opt->wim_inside_opt = 1;
break;
}
case 0 :
{
io->mod->s_opt->wim_n_rgrft = atoi(optarg);
break;
}
case 1 :
{
io->mod->s_opt->wim_n_globl = atoi(optarg);
break;
}
case 2 :
{
io->mod->s_opt->wim_max_dist = atoi(optarg);
break;
}
case 3 :
{
io->mod->s_opt->wim_n_optim = atoi(optarg);
break;
}
case 4 :
{
io->mod->s_opt->wim_n_best = atoi(optarg);
break;
}
case 16 :
{
io->mod->s_opt->min_diff_lk_local = atof(optarg);
break;
}
case 17 :
{
io->mod->s_opt->min_diff_lk_global = atof(optarg);
break;
}
case 18 :
{
io->mod->s_opt->steph_spr = 0;
io->mod->s_opt->greedy = 1;
break;
}
case 19 :
{
io->mod->s_opt->brent_it_max = atoi(optarg);
break;
}
case 20 :
{
io->mod->s_opt->random_input_tree = 1;
break;
}
case 21 :
{
io->mod->s_opt->random_input_tree = 1;
io->mod->s_opt->n_rand_starts = atoi(optarg);
if(io->mod->s_opt->n_rand_starts < 1) Exit("\n== Number of random starting trees must be > 0.\n\n");
}
case 's':case 6:
{
if((!strcmp(optarg,"spr")) || (!strcmp(optarg,"SPR")))
{
io->mod->s_opt->topo_search = SPR_MOVE;
io->mod->s_opt->greedy = (io->mod->s_opt->steph_spr)?(0):(1);
}
else if((!strcmp(optarg,"nni")) || (!strcmp(optarg,"NNI")))
{
io->mod->s_opt->topo_search = NNI_MOVE;
io->mod->s_opt->random_input_tree = 0;
}
else if((!strcmp(optarg,"best")) || (!strcmp(optarg,"BEST")))
{
io->mod->s_opt->topo_search = BEST_OF_NNI_AND_SPR;
io->mod->s_opt->greedy = (io->mod->s_opt->steph_spr)?(0):(1);
}
break;
}
case 'd':case 7:
{
if(!strcmp(optarg,"nt"))
{
io->datatype = NT;
io->mod->ns = 4;
io->mod->m4mod->n_o = 4;
if((io->mod->whichmodel == LG) ||
(io->mod->whichmodel == WAG) ||
(io->mod->whichmodel == DAYHOFF) ||
(io->mod->whichmodel == JTT) ||
(io->mod->whichmodel == BLOSUM62) ||
(io->mod->whichmodel == MTREV) ||
(io->mod->whichmodel == RTREV) ||
(io->mod->whichmodel == CPREV) ||
(io->mod->whichmodel == DCMUT) ||
(io->mod->whichmodel == VT) ||
(io->mod->whichmodel == MTMAM) ||
(io->mod->whichmodel == MTART) ||
(io->mod->whichmodel == HIVW) ||
(io->mod->whichmodel == HIVB) ||
(io->mod->whichmodel == AB) ||
(io->mod->whichmodel == CUSTOMAA)
)
{
io->mod->whichmodel = HKY85;
strcpy(io->mod->modelname->s, "HKY85\0");
}
}
else if (!strcmp(optarg,"aa"))
{
io->datatype = AA;
io->mod->s_opt->opt_kappa = NO;
io->mod->ns = 20;
io->mod->m4mod->n_o = 20;
if(
(io->mod->whichmodel == JC69) ||
(io->mod->whichmodel == K80) ||
(io->mod->whichmodel == F81) ||
(io->mod->whichmodel == HKY85) ||
(io->mod->whichmodel == F84) ||
(io->mod->whichmodel == TN93) ||
(io->mod->whichmodel == GTR) ||
(io->mod->whichmodel == CUSTOM)
)
{
io->mod->whichmodel = LG;
strcpy(io->mod->modelname->s, "LG\0");
}
}
else if ((!strcmp(optarg,"generic")) || (!strcmp(optarg,"gen")))
{
io->datatype = GENERIC;
}
else
{
char choix;
PhyML_Printf("\n. Unknown argument to -d option: please use `nt' for DNA or `aa' for Amino-Acids\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
break;
}
case 'm': case 5 :
{
int i;
For(i,strlen(optarg)) Uppercase(optarg+i);
if(!isalpha(optarg[0]))
{
strcpy(io->mod->custom_mod_string->s,optarg);
if(strlen(io->mod->custom_mod_string->s) != 6)
{
Warn_And_Exit("\n. The string should be of length 6.\n");
}
else
{
/* Make_Custom_Model(io->mod); */
/* Translate_Custom_Mod_String(io->mod); */
}
io->datatype = NT;
io->mod->whichmodel = CUSTOM;
strcpy(io->mod->modelname->s, "custom");
io->mod->s_opt->opt_kappa = NO;
io->mod->s_opt->opt_rr = YES;
}
else if (strcmp(optarg, "JC69") == 0)
{
io->datatype = NT;
io->mod->whichmodel = JC69;
}
else if(strcmp(optarg, "K80") == 0)
{
io->datatype = NT;
io->mod->whichmodel = K80;
}
else if(strcmp(optarg, "F81") == 0)
{
io->datatype = NT;
io->mod->whichmodel = F81;
}
else if (strcmp(optarg, "HKY85") == 0)
{
io->datatype = NT;
io->mod->whichmodel = HKY85;
}
else if(strcmp(optarg, "F84") == 0)
{
io->datatype = NT;
io->mod->whichmodel = F84;
}
else if (strcmp (optarg,"TN93") == 0)
{
io->datatype = NT;
io->mod->whichmodel = TN93;
}
else if(strcmp (optarg, "GTR") == 0)
{
io->datatype = NT;
io->mod->whichmodel = GTR;
}
else if(strcmp(optarg, "DAYHOFF") == 0)
{
io->datatype = AA;
io->mod->whichmodel = DAYHOFF;
}
else if(strcmp (optarg, "JTT") == 0)
{
io->datatype = AA;
io->mod->whichmodel = JTT;
}
else if(strcmp(optarg, "MTREV") == 0)
{
io->datatype = AA;
io->mod->whichmodel = MTREV;
}
else if(strcmp (optarg, "LG") == 0)
{
io->datatype = AA;
io->mod->whichmodel = LG;
}
else if(strcmp (optarg, "WAG") == 0)
{
io->datatype = AA;
io->mod->whichmodel = WAG;
}
else if(strcmp(optarg, "DCMUT") == 0)
{
io->datatype = AA;
io->mod->whichmodel = DCMUT;
}
else if(strcmp (optarg, "RTREV") == 0)
{
io->datatype = AA;
io->mod->whichmodel = RTREV;
}
else if(strcmp(optarg, "CPREV") == 0)
{
io->datatype = AA;
io->mod->whichmodel = CPREV;
}
else if(strcmp(optarg, "VT") == 0)
{
io->datatype = AA;
io->mod->whichmodel = VT;
}
else if(strcmp(optarg, "BLOSUM62") == 0)
{
io->datatype = AA;
io->mod->whichmodel = BLOSUM62;
}
else if(strcmp(optarg, "MTMAM") == 0)
{
io->datatype = AA;
io->mod->whichmodel = MTMAM;
}
else if (strcmp(optarg,"MTART") == 0)
{
io->datatype = AA;
io->mod->whichmodel = MTART;
}
else if (strcmp(optarg,"HIVW") == 0)
{
io->datatype = AA;
io->mod->whichmodel = HIVW;
}
else if(strcmp(optarg, "HIVB") == 0)
{
io->datatype = AA;
io->mod->whichmodel = HIVB;
}
else if(strcmp(optarg, "AB") == 0)
{
io->datatype = AA;
io->mod->whichmodel = AB;
}
else if (strcmp(optarg, "CUSTOM") == 0)
{
io->datatype = AA;
io->mod->whichmodel = CUSTOMAA;
}
else
{
PhyML_Printf("\n. The model name is incorrect. Please see the documentation.\n");
Exit("\n");
}
Set_Model_Name(io->mod);
break;
}
case 'a':case 14 :
{
if ((strcmp (optarg, "e") == 0) ||
(strcmp (optarg, "E") == 0) ||
(strcmp (optarg, "estimated") == 0) ||
(strcmp (optarg, "ESTIMATED") == 0))
{
io->mod->s_opt->opt_alpha = YES;
}
else if (atof(optarg) < 1.E-10)
{
char choix;
PhyML_Printf("\n. Alpha must be > 1.E-10.\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else
{
io->mod->ras->alpha->v = (phydbl)atof(optarg);
io->mod->s_opt->opt_alpha = 0;
}
break;
}
case 'b':case 10:
{
if ((int)String_To_Dbl(optarg) < -5)
{
char choix;
PhyML_Printf("\n. Branch test value must be a positive integer for bootstrap, or between -1 and -4 for aLRT branch test\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else
{
if((int)String_To_Dbl(optarg) > 0)
{
io->ratio_test = 0;
io->mod->bootstrap = (int)atoi(optarg);
io->print_boot_trees = 1;
if(io->n_data_sets > 1)
{
char choix;
PhyML_Printf("\n. Bootstrap option is not allowed with multiple data sets\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
}
else if (atoi(optarg)==0)
{
io->mod->bootstrap = 0;
io->ratio_test = 0;
}
else
{
io->mod->bootstrap = 0;
io->ratio_test = -(int)atoi(optarg);
}
}
break;
}
case 'c':case 12:
{
if ((!atoi(optarg)) || (atoi(optarg) < 0))
{
char choix;
PhyML_Printf("\n. Unknown argument to -c option: the number of rate categories must be a positive integer\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else
{
io->mod->ras->n_catg = atoi(optarg);
if(io->mod->ras->n_catg < 1)
{
PhyML_Printf("\n. The number of rate categories must be a positive integer\n");
Exit("\n");
}
}
break;
}
case 'f':
{
if(!strcmp(optarg,"e"))
{
if(io->datatype == NT)
io->mod->s_opt->opt_state_freq = NO;
else if (io->datatype == AA)
io->mod->s_opt->opt_state_freq = YES;
else
{
PhyML_Printf("\n. Please define the data type (nt or aa) before setting the -f option\n");
Exit("\n");
}
}
else if(!strcmp(optarg,"m"))
{
if (io->datatype == NT)
io->mod->s_opt->opt_state_freq = YES;
else if (io->datatype == AA)
io->mod->s_opt->opt_state_freq = NO;
else
{
PhyML_Printf("\n. Please define the data type (nt or aa) before setting the -f option\n");
Exit("\n");
}
}
else if(!isalpha(optarg[0]))
{
phydbl sum;
double val1,val2,val3,val4;
io->mod->s_opt->opt_state_freq = 0;
io->mod->s_opt->user_state_freq = 1;
/* sscanf(optarg,"%lf,%lf,%lf,%lf", */
/* io->mod->user_b_freq, */
/* io->mod->user_b_freq+1, */
/* io->mod->user_b_freq+2, */
/* io->mod->user_b_freq+3); */
sscanf(optarg,"%lf,%lf,%lf,%lf",&val1,&val2,&val3,&val4);
io->mod->user_b_freq->v[0] = (phydbl)val1;
io->mod->user_b_freq->v[1] = (phydbl)val2;
io->mod->user_b_freq->v[2] = (phydbl)val3;
io->mod->user_b_freq->v[3] = (phydbl)val4;
sum =
(io->mod->user_b_freq->v[0] +
io->mod->user_b_freq->v[1] +
io->mod->user_b_freq->v[2] +
io->mod->user_b_freq->v[3]);
io->mod->user_b_freq->v[0] /= sum;
io->mod->user_b_freq->v[1] /= sum;
io->mod->user_b_freq->v[2] /= sum;
io->mod->user_b_freq->v[3] /= sum;
if(io->mod->user_b_freq->v[0] < .0 ||
io->mod->user_b_freq->v[1] < .0 ||
io->mod->user_b_freq->v[2] < .0 ||
io->mod->user_b_freq->v[3] < .0 ||
io->mod->user_b_freq->v[0] > 1. ||
io->mod->user_b_freq->v[1] > 1. ||
io->mod->user_b_freq->v[2] > 1. ||
io->mod->user_b_freq->v[3] > 1.)
{
Warn_And_Exit("\n. Invalid base frequencies.\n");
}
}
break;
}
case 'h':case 69:
{
Usage();
break;
}
case 'i':case 9:
{
char *tmp;
tmp = (char *) mCalloc (T_MAX_FILE, sizeof(char));
if (strlen (optarg) > T_MAX_FILE -16)
{
char choix;
strcpy (tmp, "\n. The file name'");
strcat (tmp, optarg);
strcat (tmp, "' is too long.\n");
PhyML_Printf("%s",tmp);
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else if (!Filexists (optarg))
{
char choix;
strcpy (tmp, "\n. The file '");
strcat (tmp, optarg);
strcat (tmp, "' does not exist.\n");
PhyML_Printf("%s",tmp);
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else
{
strcpy(io->in_align_file, optarg);
io->fp_in_align = Openfile(io->in_align_file,0);
strcpy(io->out_file, optarg);
strcpy(io->out_tree_file,optarg);
#ifdef PHYML
strcat(io->out_tree_file,"_phyml_tree");
#elif M4
strcat(io->out_tree_file,"_m4_tree");
#elif PHYREX
strcat(io->out_tree_file,"_phyrex_tree");
#endif
strcpy(io->out_stats_file,optarg);
#ifdef PHYML
strcat(io->out_stats_file,"_phyml_stats");
#elif M4
strcat(io->out_stats_file,"_m4_stats");
#elif PHYREX
strcat(io->out_stats_file,"_phyrex_stats");
#endif
#ifdef PHYREX
strcpy(io->out_summary_file,optarg);
strcat(io->out_summary_file,"_phyrex_summary");
#endif
}
Free (tmp);
break;
}
case 't':case 11:
{
if ((io->mod->whichmodel != JC69) &&
(io->mod->whichmodel != F81) &&
(io->mod->whichmodel != GTR))
{
if ((strcmp(optarg, "e") == 0) ||
(strcmp(optarg, "E") == 0) ||
(strcmp(optarg, "estimated") == 0) ||
(strcmp(optarg, "ESTIMATED") == 0))
{
io->mod->kappa->v = 4.0;
io->mod->s_opt->opt_kappa = YES;
if (io->mod->whichmodel == TN93)
io->mod->s_opt->opt_lambda = YES;
}
else
{
if (atof(optarg) < .0)
{
char choix;
PhyML_Printf("\n. The ts/tv ratio must be a positive number\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else
{
io->mod->kappa->v = (phydbl)atof(optarg);
io->mod->s_opt->opt_kappa = 0;
io->mod->s_opt->opt_lambda = 0;
}
}
}
break;
}
case 'n':case 8:
{
if ((!atoi(optarg)) || (atoi(optarg) < 0))
{
char choix;
PhyML_Printf("\n. The number of alignments must be a positive integer\n");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else io->n_data_sets = atoi (optarg);
break;
}
case 'q':case 22:
{
io->interleaved = NO;
break;
}
case 'u':case 15:
{
char *tmp;
tmp = (char *)mCalloc(T_MAX_FILE, sizeof(char));
if(strlen(optarg) > T_MAX_FILE -11)
{
char choix;
strcpy (tmp, "\n. The file name'");
strcat (tmp, optarg);
strcat (tmp, "' is too long.\n");
PhyML_Printf("%s",tmp);
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else if (! Filexists (optarg))
{
char choix;
strcpy (tmp, "\n. The file '");
strcat (tmp, optarg);
strcat (tmp, "' doesn't exist.\n");
PhyML_Printf("%s",tmp);
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else
{
strcpy(io->in_tree_file, optarg);
io->in_tree = 2;
io->fp_in_tree = Openfile(io->in_tree_file,READ);
}
Free(tmp);
break;
}
case 'v':case 13:
{
if ((strcmp (optarg, "e") == 0) ||
(strcmp (optarg, "E") == 0) ||
(strcmp (optarg, "estimated") == 0) ||
(strcmp (optarg, "ESTIMATED") == 0))
{
io->mod->s_opt->opt_pinvar = YES;
io->mod->ras->invar = YES;
}
else if ((atof(optarg) < 0.0) || (atof(optarg) > 1.0))
{
char choix;
PhyML_Printf("\n. The proportion of invariable site must be a number between 0.0 and 1.0\n");
PhyML_Printf("\n. Type any key to exit.");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
else
{
io->mod->ras->pinvar->v = (phydbl)atof(optarg);
if (io->mod->ras->pinvar->v > 0.0+SMALL)
io->mod->ras->invar = 1;
else
io->mod->ras->invar = 0;
io->mod->s_opt->opt_pinvar = 0;
}
break;
}
case 'o':
{
if(!strcmp(optarg,"tlr"))
{
io->mod->s_opt->opt_topo = YES;
io->mod->s_opt->opt_bl = YES;
io->mod->s_opt->opt_subst_param = YES;
}
else if(!strcmp(optarg,"tl"))
{
io->mod->s_opt->opt_topo = YES;
io->mod->s_opt->opt_bl = YES;
io->mod->s_opt->opt_subst_param = NO;
}
else if(!strcmp(optarg,"t"))
{
Warn_And_Exit("\n. You can't optimize the topology without adjusting branch length too...\n");
}
else if(!strcmp(optarg,"lr"))
{
io->mod->s_opt->opt_topo = NO;
io->mod->s_opt->opt_bl = YES;
io->mod->s_opt->opt_subst_param = YES;
}
else if(!strcmp(optarg,"l"))
{
io->mod->s_opt->opt_topo = NO;
io->mod->s_opt->opt_bl = YES;
io->mod->s_opt->opt_subst_param = NO;
}
else if(!strcmp(optarg,"r"))
{
io->mod->s_opt->opt_topo = NO;
io->mod->s_opt->opt_bl = NO;
io->mod->s_opt->opt_subst_param = YES;
}
else if(!strcmp(optarg,"none") || !strcmp(optarg,"n"))
{
io->mod->s_opt->opt_topo = NO;
io->mod->s_opt->opt_bl = NO;
io->mod->s_opt->opt_subst_param = NO;
}
else
{
char choix;
PhyML_Printf ("\n. The optimization parameter must be 'tlr' or 'tl' or 'lr' or 'l' or 'r' or ''.");
PhyML_Printf("\n. Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
break;
}
case '?':
{
Exit("\n");
break;
}
case -1:
{
break;
}
default:
{
Usage();
break;
}
}
}while(c != -1);
/* if((io->mod->whichmodel == K80) || (io->mod->whichmodel == JC69)) */
/* { */
/* if(io->mod->s_opt->opt_state_freq) */
/* { */
/* char c; */
/* PhyML_Printf("\n. WARNING: nucleotide frequencies must be set to 1/4 with this model.\n"); */
/* PhyML_Printf("\n. Type the enter key to resume the analysis.\n"); */
/* scanf("%c",&c); */
/* } */
/* io->mod->s_opt->opt_state_freq = 0; */
/* } */
if(io->mod->s_opt->constrained_br_len == YES)
{
io->mod->s_opt->opt_topo = NO;
/* io->mod->s_opt->opt_bl = NO; */
}
#ifndef PHYML
if((open_ps_file) || (io->m4_model == YES))
{
strcpy(io->out_ps_file,io->in_align_file);
strcat(io->out_ps_file, "_mc_tree.ps");
io->fp_out_ps = Openfile(io->out_ps_file,WRITE);
}
#endif
if(io->datatype == UNDEFINED) io->datatype = NT;
if((io->mod->s_opt->n_rand_starts) &&
(io->mod->s_opt->topo_search == NNI_MOVE) &&
(io->mod->s_opt->random_input_tree))
{
Warn_And_Exit("\n== The random starting tree option is only compatible with SPR based search options.\n");
}
if ((io->datatype == NT) && (io->mod->whichmodel > 10))
{
char choix;
PhyML_Printf("\n== Err.: model incompatible with the data type. Please use JC69, K80, F81, HKY, F84, TN93 or GTR\n");
PhyML_Printf("\n== Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Warn_And_Exit("\n");
}
else if ((io->datatype == AA) && (io->mod->whichmodel < 11))
{
char choix;
PhyML_Printf("\n== Err.: model incompatible with the data type. Please use LG, Dayhoff, JTT, MtREV, WAG, DCMut, RtREV, CpREV, VT, Blosum62, MtMam, MtArt, HIVw, HIVb or AB.\n");
PhyML_Printf("\n== Type any key to exit.\n");
if(!scanf("%c",&choix)) Exit("\n");
Exit("\n");
}
if(io->mod->use_m4mod == NO)
{
io->mod->s_opt->opt_cov_delta = 0;
io->mod->s_opt->opt_cov_alpha = 0;
io->mod->s_opt->opt_cov_free_rates = 0;
}
if((io->mod->s_opt->opt_cov_free_rates) && (io->mod->s_opt->opt_cov_alpha))
{
io->mod->s_opt->opt_cov_free_rates = 1;
io->mod->m4mod->use_cov_alpha = 0;
io->mod->m4mod->use_cov_free = 1;
}
if(io->print_site_lnl)
{
strcpy(io->out_lk_file,io->in_align_file);
strcat(io->out_lk_file, "_phyml_lk");
if(io->append_run_ID) { strcat(io->out_lk_file,"_"); strcat(io->out_lk_file,io->run_id_string); }
io->fp_out_lk = Openfile(io->out_lk_file,1);
}
if(io->print_trace)
{
strcpy(io->out_trace_file,io->in_align_file);
strcat(io->out_trace_file,"_phyml_trace");
if(io->append_run_ID) { strcat(io->out_trace_file,"_"); strcat(io->out_trace_file,io->run_id_string); }
io->fp_out_trace = Openfile(io->out_trace_file,1);
}
if(io->mod->s_opt->random_input_tree)
{
strcpy(io->out_trees_file,io->in_align_file);
strcat(io->out_trees_file,"_phyml_rand_trees");
if(io->append_run_ID) { strcat(io->out_trees_file,"_"); strcat(io->out_trees_file,io->run_id_string); }
io->fp_out_trees = Openfile(io->out_trees_file,1);
}
if((io->print_boot_trees) && (io->mod->bootstrap > 0))
{
strcpy(io->out_boot_tree_file,io->in_align_file);
strcat(io->out_boot_tree_file,"_phyml_boot_trees");
if(io->append_run_ID) { strcat(io->out_boot_tree_file,"_"); strcat(io->out_boot_tree_file,io->run_id_string); }
io->fp_out_boot_tree = Openfile(io->out_boot_tree_file,1);
strcpy(io->out_boot_stats_file,io->in_align_file);
strcat(io->out_boot_stats_file,"_phyml_boot_stats");
if(io->append_run_ID) { strcat(io->out_boot_stats_file,"_"); strcat(io->out_boot_stats_file,io->run_id_string); }
io->fp_out_boot_stats = Openfile(io->out_boot_stats_file,1);
}
if(io->append_run_ID)
{
strcat(io->out_tree_file,"_");
strcat(io->out_stats_file,"_");
strcat(io->out_tree_file,io->run_id_string);
strcat(io->out_stats_file,io->run_id_string);
}
if(io->mod->ras->n_catg == 1) io->mod->s_opt->opt_alpha = 0;
if(!io->mod->s_opt->opt_subst_param)
{
io->mod->s_opt->opt_alpha = 0;
io->mod->s_opt->opt_kappa = 0;
io->mod->s_opt->opt_lambda = 0;
io->mod->s_opt->opt_pinvar = 0;
io->mod->s_opt->opt_rr = 0;
}
if(io->mod->whichmodel != K80 &&
io->mod->whichmodel != HKY85 &&
io->mod->whichmodel != F84 &&
io->mod->whichmodel != TN93)
{
io->mod->s_opt->opt_kappa = 0;
}
if(io->datatype == AA && io->mod->whichmodel == CUSTOMAA && !io->mod->fp_aa_rate_mat)
{
PhyML_Printf("\n== Custom model option with amino-acid requires you to specify a rate matrix file through the '--aa_rate_file' option.\n");
Exit("\n");
}
#if !defined(PHYTIME)
// Make sure you don't erase the input file...
if(!strcmp(io->out_tree_file,io->in_align_file) ||
!strcmp(io->out_stats_file,io->in_align_file)) Generic_Exit(__FILE__,__LINE__,__FUNCTION__);
writemode = WRITE;
io->fp_out_tree = Openfile(io->out_tree_file,writemode);
io->fp_out_stats = Openfile(io->out_stats_file,writemode);
#endif
#if defined(PHYREX)
io->fp_out_summary = Openfile(io->out_summary_file,writemode);
#endif
writemode++; // just to silence a warning message at compilation
if(io->mod->whichmodel == GTR)
{
/* Make_Custom_Model(io->mod); */
io->mod->s_opt->opt_rr = 1;
}
return 1;
}
void bNNI (tree *T, double **avgDistArray, int *count, FILE *statfile)
{
edge *e;
edge **edgeArray;
int *p, *location, *q;
int i;
int possibleSwaps;
double *weights;
p = initPerm (T->size+1);
q = initPerm (T->size+1);
edgeArray = (edge **) mCalloc ((T->size+1), sizeof (edge *));
weights = (double *) mCalloc ((T->size+1), sizeof (double));
location = (int *) mCalloc ((T->size+1), sizeof (int));
for (i=0; i<T->size+1; i++)
{
weights[i] = 0.0;
location[i] = NONE;
}
assignBMEWeights (T, avgDistArray);
weighTree (T);
if (!isBoostrap)
{
if (statfile)
fprintf (statfile, "\tBefore NNI: tree length is %f.\n", T->weight);
if (verbose > 2)
Debug ( (char*)"Before NNI: tree length is %f.", T->weight);
else if (verbose > 1)
Message ( (char*)". Before NNI: tree length is %f.", T->weight);
}
e = findBottomLeft (T->root->leftEdge);
while (NULL != e)
{
edgeArray[e->head->index+1] = e;
location[e->head->index+1] = bNNIEdgeTest (e, T, avgDistArray,
weights + e->head->index + 1);
e = depthFirstTraverse (T,e);
}
possibleSwaps = makeThreshHeap (p, q, weights, T->size+1,0.0);
permInverse (p, q, T->size+1);
/* We put the negative values of weights into a heap, indexed by p
* with the minimum value pointed to by p[1]
* p[i] is index (in edgeArray) of edge with i-th position in the
* heap, q[j] is the position of edge j in the heap */
while (weights[p[1]] < -DBL_EPSILON)
{
(*count)++;
T->weight = T->weight + weights[p[1]];
if (!isBoostrap)
{
if (statfile)
fprintf (statfile, "\tNNI %5d: new tree length is %f.\n", *count, T->weight);
if (verbose > 2)
Debug ( (char*)"NNI %5d: new tree length is %f.", *count, T->weight);
else if (verbose > 1)
Message ( (char*)". NNI %5d: new tree length is %f.", *count, T->weight);
}
bNNItopSwitch (edgeArray[p[1]], location[p[1]], avgDistArray);
location[p[1]] = NONE;
weights[p[1]] = 0.0; //after the bNNI, this edge is in optimal configuration
popHeap (p, q, weights, possibleSwaps--, 1);
/* but we must retest the other edges of T */
/* CHANGE 2/28/2003 expanding retesting to _all_ edges of T */
e = depthFirstTraverse (T, NULL);
while (NULL != e)
{
bNNIRetestEdge (p, q, e, T, avgDistArray, weights, location, &possibleSwaps);
e = depthFirstTraverse (T, e);
}
}
free (p);
free (q);
free (location);
free (edgeArray);
free (weights);
assignBMEWeights (T, avgDistArray);
return;
}
