meme.h

EM算法的改进

 				   	0 : no palindromes
                                   	1 : force DNA palindromes
				*/
  THETA map;			/* letter to frequency vector mapping */
  THETA lomap;			/* letter to logodds vector mapping */
  MOTIF motifs[NMOTIFS];	/* known motifs in dataset */
  int nkmotifs;			/* number of known motifs in dataset */
  NEGTYPE negtype;		/* how to use negative examples */
  int back_order;		/* order of Markov background model */
  double *back;			/* Markov background model: 
				   back[s2i(wa)] = log Pr(a | wa) */
  double log_total_prob;	/* total (log) cumulative background prob. */
  PRIORS *priors;		/* the prior probabilities model */
  P_POINT *p_point;		/* previously learned starting points */
  double wnsites;		/* weight on prior on nsites */
  BOOLEAN ma_adj;		/* adjust width/pos. using mult. algn. method */
  double wg;			/* gap cost (initialization) */
  double ws;			/* space cost (extension) */
  BOOLEAN endgaps;		/* penalize end gaps if TRUE */
  double distance;		/* convergence radius */
  double prob;			/* sampling probability for subsq. starts */
  int nmotifs;			/* number of motifs to find */
  int maxiter;			/* maximum number of iterations for EM */
  double evt;			/* E-value threshold */
  char *mod;			/* name of model */
  char *mapname;		/* name of spmap */
  double map_scale;		/* scale of spmap */
  char *priorname;		/* name of type of prior */
  double beta;			/* beta for prior */
  int seed;			/* random seed */
  double seqfrac;		/* fraction of sequences to use */
  char *plib_name;		/* name of file with prior library */
  char *bfile;			/* name of background model file */
  char *datafile;		/* name of the dataset file */
  char *negfile;		/* name of negative examples file */
  char *command;		/* command line */
  OBJTYPE objfun;		/* objective function */
  THETA pairwise;		/* contains score matrix for pairwise scores */
  double min_ic;		/* min per-column information content */
  char *meme_directory;		/* meme source directory */
  double max_time;		/* maximum allowed CPU time */
} DATASET;

/* motif occurrence in sequence */
typedef struct {
  int seqno;		/* index in samples array */
  int pos;		/* character position of site in sequence */
  double zij;		/* value of z_ij */
  int invcomp;		/* on inverse complement strand */
} SITE;

/* tiling of sequence with motifs */
typedef struct {
  int *hits;		/* hit[i] = m, motif m occurs at position i in seq 
			   <0, on reverse strand, =0 means no hit */
  double *pvalues;	/* pvalues[i] is p-value of match at position i */
  double pv;		/* p-value of product of p-values of best hits */
  char *diagram;	/* block diagram for printing */
} TILING;

/* subroutines */

extern double exp(double x);
extern int em_subseq(
  THETA map,                    /* freq x letter map */
  DATASET *dataset,             /* the dataset */
  MODEL *model,                 /* the model */
  PRIORS *priors,               /* the priors */
  int w,                        /* width of motif */
  int n_nsites0,                /* number of nsites0 values to try */
  double alpha,                 /* sampling probability */
  P_POINT *p_point,             /* starting point for previous components */
  S_POINT s_points[]            /* array of starting points: 1 per nsites0 */
);
extern int subseq7(
  MTYPE mtype,			/* type of model */
  BOOLEAN revcomp,              /* use reverse complement strand of DNA */
  THETA map,			/* freq x letter map */
  DATASET *dataset,		/* the dataset */
  int w,			/* width of motif */
  int n_nsites0,		/* number of nsites0 values to try */
  S_POINT s_points[]  		/* array of starting points: 1 per nsites0 */
);
extern int pY_compare(
  const void *v1,
  const void *v2
);
extern void get_not_o(
  DATASET *dataset,			/* the dataset */
  int w,				/* width of motif */
  BOOLEAN get_log			/* compute log_not_o if true */
);
extern double get_log_sig(
  double llr,					/* log likelihood ratio */
  MTYPE mtype,					/* type of model */
  int w,					/* width of motif */
  double wN,					/* weighted number of sites */
  int N,					/* number of sites */
  BOOLEAN invcomp,				/* inv. compl. strand, too */
  BOOLEAN pal, 					/* motif is DNA palindrome */
  DATASET *dataset				/* the dataset */
);
extern void calc_entropy(
  MODEL *model,			/* the model */
  DATASET *dataset  		/* the dataset */
);
extern double log_comb(
  int m,				/* length of sequence */
  int n 				/* number of segments */
);
extern double get_log_nalign(
  MTYPE mtype,					/* type of model */
  int w, 					/* width of motif */
  int N,					/* number of occurrences */
  BOOLEAN invcomp,                              /* inv. compl. seq allowed */
  DATASET *dataset  				/* the dataset */
);
extern void adjust_motif(
  MODEL *model,				/* the model */
  MODEL *scratch_model,			/* the scratch model */
  DATASET *dataset,			/* the dataset */
  PRIORS *priors,			/* the priors */
  double wnsites,			/* weight on prior on nsites */
  BOOLEAN ma_adj,			/* adjust w using mult. algn. method  */
  BOOLEAN palindrome,			/* convert motif to palindrome */
  int c,				/* component of model to adjust */
  int min_w,				/* minimum width of motif allowed */
  int max_w,				/* maximum width of motif allowed */
  int maxiter,				/* maximum number iterations for EM */
  double distance,			/* stopping criterion */
  double wg,				/* gap cost (initialization) */
  double ws,				/* space cost (extension) */
  BOOLEAN endgaps 			/* penalize end gaps if TRUE */
);
extern void init_theta(
  THETA theta,			/* theta */
  char *start,			/* integer encoded starting sequence */
  int w,			/* width of motif */
  THETA map,			/* frequency vectors for each letter */ 
  int alength			/* alphabet length */
);
extern S_POINT *get_starts(
  DATASET *dataset,		/* the dataset */
  MODEL *model,                 /* the model */
  char *e_cons,			/* encoded consensus sequence */
  double w_factor,		/* factor between sampled widths */
  int *n_starts			/* number of starting points */
);
extern THETA init_map(
  MAP_TYPE type,		/* type of mapping:
					Uni	- add n prior
					Pam	- pam matrix
				*/
  double scale,			/* degree of crispness, depends on type,
					Uni	- add n prior (n)
					Pam	- pam distance
				*/
  int alength,			/* length of alphabet */
  double *back,			/* background frequencies */
  BOOLEAN lo			/* create logodds matrix */
);
extern void copy_theta(
  THETA s,	 		/* source */
  THETA d,			/* destination */
  int w,			/* width of motif */
  int alength			/* length of alphabet */
);
extern void copy_model(
  MODEL *m1, 				/* source */
  MODEL *m2,				/* destination */
  int alength				/* length of alphabet */
);
extern void init_meme(
  int argc,			/* number of input arguments */
  char **argv,			/* input arguments */
  MODEL **model_p,      	/* the model */
  MODEL **best_model_p, 	/* the best model */
  MODEL **scratch_model_p, 	/* the best model */
  MODEL **neg_model_p,  	/* model of negative examples */
  DATASET **dataset_p,  	/* the dataset */
  DATASET **neg_dataset_p       /* dataset of negative examples */
);
extern MODEL *create_model(
  MTYPE mtype,				/* type of model */
  BOOLEAN invcomp,			/* use inv comp strand  */
  int max_w,				/* maximum width of motif */
  int alength				/* length of alphabet */
);
extern double min_sites(
  double nu,			/* degrees of freedom */
  double alpha,			/* significance level */
  double max_h			/* maximum entropy */
);

extern int read_motifs (
  FILE *fdata,                          /* opened dataset file */
  char *filename,                       /* motif file */
  MOTIF motif[NMOTIFS],                 /* motif info */
  BOOLEAN save_dataset,                 /* return dataset in memory */
  DATASET *dataset                      /* integer-encoded dataset */
);

extern SITE *get_sites(
  DATASET *dataset,			/* the dataset */
  MODEL *model,				/* the model */
  int *n,				/* number of sites found */
  int *best_site			/* index of best site in array */
);

extern LOGODDS make_log_odds(
  THETA theta1,			/* motif theta */
  THETA theta0,			/* negative theta; use 0 if NULL */
  double *back,			/* background frequencies; use 0 if NULL */
  double q,			/* mixing parameter for theta0 and back */
  int w,			/* width of motif */
  int alength 			/* length of alphabet */
);

extern int get_max(
  MTYPE mtype,		/* the model type */
  DATASET *dataset,	/* the dataset */
  int w,		/* length of sites */ 
  P_PROB maxima, 	/* array of encoded site starts of local maxima */
  BOOLEAN ic, 		/* use inverse complement, too */
  BOOLEAN sort		/* sort the maxima */
);

extern int align_top_subsequences(
  MTYPE mtype,				/* type of model */
  int w,				/* width of motif */
  DATASET *dataset,			/* the dataset */
  int iseq,				/* sequence number of starting point */
  int ioff,				/* sequence offset of starting point */
  char *eseq,				/* integer encoded subsequence */
  char *name,				/* name of sequence */
  int n_nsites0,			/* number of nsites0 values to try */
  int n_maxima,				/* number of local maxima */
  P_PROB maxima,			/* sorted local maxima indices */
  double *col_scores,			/* column scores for last start point */
  S_POINT s_points[]			/* array of starting points */
);

extern double log_qfast(
  int n,			/* number of random variables in product */
  double logk			/* product of random variables */
);

#include <star.h>
#include <llr.h>
#include <em.h>
#include <hash_alph.h>
#include <read_seq_file.h>
#include <display.h>
#include <dpalign.h>
#include <histogram.h>
#include <clock.h>
#include <message.h>

#endif