/*------------------------------------------------------------------------*/
/*(C) Copyright 1991-2006, Fred Hutchinson Cancer Research Center         */
/*      motifj.h  Header file for PROTOMAT programs                       */
/* NOTE for Silicon Graphics users:  The type of scores in
       struct score should be changed from char to int to get correct
       processing (but not for SUN!)                                      */
/*------------------------------------------------------------------------*/
/*  6/29/90 J. Henikoff
    1/28/99 Increased SNAMELEN from 11 to 18; IDLEN from 10 to 12
    2/21/00 Added id->full_entry to struct db_id
    8/20/01 Increased MAXSEQS and MAXFREQ from 400 to 600
    1/ 2/06 Increase MAXSEQS and MAXFREQ from 600 to 1000
   12/23/06 Increased SNAMELEN from 18 to 20
--------------------------------------------------------------------------*/
#include <stdlib.h>
#include <stdio.h>
#include <string.h>

#define VERSION		   8		/*  motifj version number */
#define YES                1
#define NO                 0
#define ESC               27
#define CR                13
#define LF                10
#define UMIN(x, y)	  ( (x<y) ? x : y)   /* UNIX min macro */
#define UMAX(x, y)	  ( (x>y) ? x : y)   /* UNIX max macro */
/*   INDEX & INDEXCOL compute the sequential indices for the lower half of an
     nxn symmetric matrix given row & column coordinates.  Lower half has
     n(n-1)/2 entries; col=0,n-2 and row=col+1,n-1; col has n-col-1 rows.
     Index runs from 0 to n(n-1)/2 - 1 down columns with
     (index=0)==(col=0,row=1) and (index=n(n-1)/2-1)==(col=n-2,row=n-1).  */
#define INDEXCOL(n, col)    ( col*n - (col*(col+1))/2 )
#define INDEXCOLROW(n, col, row)  ( col*n - (col*(col+3))/2 - 1 + row )

#define randomize()       srand((unsigned)time(NULL))  /* Seed rand() */

#define MAX_DISTANCE  	  24	/* Max spacing between aminos of motif */
#define MIN_DISTANCE      1     /* Min distance specification */
#define MAXSEQS	  	  1000  /* Max number of sequences to be analyzed */
#define MINSEQS           2     /* Min number of sequences to be analyzed */
#define MAXFREQ           1000  /* Max occurences of motif in all seqs */
#define MAX_LENGTH	  5500  /* Max length of each sequence */
#define MIN_DOMAIN_WIDTH  10	/* Minimum width */
#define MAX_DOMAIN_WIDTH  55	/* Maximum width */
#define MAX_MERGE_WIDTH   55	/* Max. width of merged blocks */
#define RELEVANT_MOTIFS   50	/* Only top scoring motifs are retained */
#define MAX_MOTIFS	  100	/* Buffer motifs before discarding */
#define MINSCORE          1     /* Min block trimming column score (0-2500)*/
#define CLTHRES            80   /* Clustering identity percentage (0-100)*/
#define DROPSCORE         -10   /* Default std devs *10 for dropping block */
#define MOTAUTO4            3   /* max. # motifs for run type 4  */
#define MOTAUTO3	    6   /* min. # motifs for run type 3 */
#define MAXBLK             15   /* max # blocks for shotgun assembly */
#define MAXTITLE	   75   /* max sequence title length */

#define PROTEIN_SUBDIRECTORY "pros/"  /* Subdirectory containing proteins */
#define PROTEIN_EXTENSION    ".pro"    /* Extension for all protein files */
#define READ                 "r"       /* Code to read disk files */
#define SNAMELEN              20       /* Max length of sequence name */
#define IDLEN                 12       /* Max length of db id */
#define FNAMELEN             160       /* Max length of file name */
#define MAXLINE               480      /* Max line length for ASCII file */
#define MATSIZE               21       /* Scoring matrix dimension */
#define HIGHPASS              4	       /* Default high pass filter value */

#define round(x) ((x >= 0.0) ? (int) (x+0.5) : (int) (x-0.5))

/* Declare new data types */
typedef unsigned char *aa_type[20][20][MAX_DISTANCE];

/* Structure to store information about each motif: */
/*  NOTE: integer fields defined as unsigned char to save space;
	  they must not exceed 255 in value */
struct motif_struct {
  unsigned char aa1, aa2, aa3, distance1, distance2;
  /*  freq is the number of sequences with this motif */
  int freq, dups;
  /*   seq_no[freq] lists the sequence numbers that have the motif,
       pos[freq] lists the offset of the motif in the corresponding
       sequences; so pos[x] is the offset into sequence # seq_no[x], 
       NOT into sequence # x */
  int seq_no[MAXFREQ], pos[MAXFREQ];
  int score, scores[MAX_DOMAIN_WIDTH], domain, mots;
  char group, sub_group;
  };

/* Structure to store information about groups for motif map routine: */
struct group_struct {
  int group_no, sub_no, position;
  };

/*-------------------------------------------------------------------*/
/*  merged_motif is an array of motifs.  Each entry is one or more   */
/*    motifs.  Each can be thought of as a block of sequences aligned*/
/*    around all the motifs.                                         */
/*-------------------------------------------------------------------*/
struct merged_motif {
	int dropped;			/* YES if dropped */
	char aa[3];			/* Amino acid motif */
	int nmotif;			/* number of motifs, >= 0  */
					/* 0 => merged (inactive) */
	int nident;			/* number of identities */
					/* ..occurs >=SIGNIF in a col. */
	int max_score;			/* max. score of merged motifs */
	int domain;			/* displacement of merged motif blocks */
					/* block width = domain+1 */
	int distance;			/* width of motifs within block */
	int dups;			/* total # dups in all seqs */
	int loffset;			/* 1st position of motif within block*/
	int leftpos[MAXSEQS];		/* leftmost position of motif   */
					/*... for each sequence*/
	int cluster[MAXSEQS];		/* cluster number for each seq */
	int scores[MAX_MERGE_WIDTH];	/* column scores */
	int t_loffset;			/* 1st position of motif within */
					/*  trimmed block */
	int t_domain;			/* trimmed block width=t_domain+1*/
	int t_score;			/* score over trimmed block */
	int position[MAXSEQS];		/* position of block within each seq*/
	int maxpos;			/* max position in any seq */
	int minpos;			/* min position in any seq */
	int in_degree;			/* in-degree of block in DAG */
	int out_degree;			/* out-degree of block in DAG */
};

/*------------------------------------------------------------------------*/
/*   sequences contains all information about the sequences.              */
/*------------------------------------------------------------------------*/
struct sequences {
	int num;		/* number of sequences */
	int totlen;		/* total length of all sequences */
	int *len;		/* lengths of each sequence */
	int *offlen;		/* offset to start of each sequence */
	char *name;		/* 10 char name of each sequence */
	char *seq;		/* sequence bases */
};

/*------------------------------------------------------------------------*/
/*   aux_seq is a list of blocks ordered left to right within a sequence  */
/*  Each sequence has the same number of blocks, but the blocks may be
     arranged in different orders in each sequence.                       */
/*------------------------------------------------------------------------*/
struct aux_seq {
	int block[RELEVANT_MOTIFS];	/* index of block in each position*/
};

struct temp {			/* temporary structure for sorting */
	int value;
	int index;
	int flag;
};
struct dtemp {			/* temporary structure for sorting */
	double value;
	int index;
};

/*-----------------------------------------------------------------------*/
/*    Structure for pairs of sequences.                                  */
/*     pair should be allocated as an array, & the number of the         */
/*     sequences forming the pair inferred from the array index.         */
/*-----------------------------------------------------------------------*/
struct pair {
	int score;		/* # of identities within trimmed block */
	int cluster;		/* cluster # for this pair */
};

/*-----------------------------------------------------------------------*/
/*  block_list is a list of blocks in an order that is consistent among
    all sequences;  it implies a partial multiple alignment.
    The next_block list is a list of all blocks, some of which may
    overlap.  The doubly linked next_best/prev_best list is a subset
    consisting of the best non-overlapping blocks in the list.           */
/*-----------------------------------------------------------------------*/
struct block_list {
	int b;			/* index of block (merged_motif) */
	int minprev;		/* min. distance from previous block */
	int maxprev;		/* max. distance from previous block */
	struct block_list *next_block;  /* all blocks in the list  */
	struct block_list *next_best;   /* best blocks in the list */
	struct block_list *prev_best;
};

/*------------------------------------------------------------------------*/
/*  path is a list of all possible paths through the blocks in all seqs.
     The first_block list includes all blocks, including those that
     possible overlap.  The first_best list includes the best sub-path
     of non-overlapping blocks.                                           */
/*------------------------------------------------------------------------*/
struct path {
	int nblocks;		/* # of blocks in path */
	int nbest;		/* # of blocks in best sub-path */
	int naas;		/* # of AAs in best sub-path */
	unsigned long totscore;	/* sum of scores of blocks in best sub-path*/
	int totmotif;		/* # motifs in best sub-path*/
	int totident;		/* # conserved residues in best sub_path */
	int nseqs;		/* # seqs in best sub-path */
	int seqs[MAXSEQS];	/* YES if path holds for sequence */
	struct block_list *first_block;  /* first block in path */
	struct block_list *first_best;	 /* first block in best sub-path*/

	struct path *next_path;
};

/*----------------------------------------------------------------------*/
/*  Ajacency matrix representation of distances between blocks in
    all sequences.                                                      */
/*----------------------------------------------------------------------*/
struct matrix {
	int npos;		/* # of seqs with positive diff in cell */
	int maxdiff;		/* maximum positive difference in cell  */
/*	int dist[MAXSEQS];	 distance from row to col for seq s */
	int mark;		/* all-purpose flag */
};

struct follow_data {
	struct path *path;
	int mark[RELEVANT_MOTIFS][RELEVANT_MOTIFS];
};

/* -------- Scoring matrix structure ----------------------------------*/
struct score {
	char scores[MATSIZE][MATSIZE];	/* valid range -127 to +128 */
	int highpass;			/* high pass filter value */
};

/*-Structure to split up a file name of the form <directory>\<name>.ext -*/
struct split_name {
	int dir_len, file_len, name_len;
};
/*------ Structure to hold the contents of a .lis or .lst file ------*/
struct db_id {
   char entry[SNAMELEN+1];	/* sequence name */
   char full_entry[2*SNAMELEN];	/* enhanced sequence name */
   char ps[2];			/* PS type=T, F or P */
   char info[FNAMELEN];		/* additional text info */
   int len;			/* sequence length */
   int frag;			/* YES if seq is a fragment */
   int lst;			/* seq in .lst file */
   int found;			/* seq found in database */
   int block;			/* seq found in block */
   int search;			/* used by excluded.c => use seq for search*/
   int rank;			/* used by matodat.c, fastodat.c */
   int score;			/* used by matodat.c, fastodat.c */
   double pvalue;		/* P-value */
   struct db_id *next;
   struct db_id *prior;
};

struct db_id *makedbid();
struct db_id *check_entry();
int get_ids();

struct split_name *split_names();

char *dir_unix();
int kr_atoi();
void kr_itoa();

void getscore();

char *num_to_aachar();
int aachar_to_num();
void pr_num_to_aa();
void pr_num_to_aa_space();