BGI/SOAPdenovo2/standardPregraph/inc/global.h

   1 /*
   2  * inc/global.h
   3  *
   4  * Copyright (c) 2008-2012 BGI-Shenzhen <soap at genomics dot org dot cn>.
   5  *
   6  * This file is part of SOAPdenovo.
   7  *
   8  * SOAPdenovo is free software: you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation, either version 3 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * SOAPdenovo is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License
  19  * along with SOAPdenovo.  If not, see <http://www.gnu.org/licenses/>.
  20  *
  21  */
  22
  23 int visual = 0;         // 1 for output some files , which are useful for visual
  24 int * contig_index_array = NULL;
  25 int scaffNum = 0;
  26 int gapNum = 1;
  27 boolean fill = 0; // 1 for output some files ,which are useful for the software "kgf"
  28 int overlaplen = 23;//k-mer Size
  29 int inGraph;        //for checking whether -g  is set, (graph prefix)
  30 long long n_ban;    //not used
  31 long long n_solexa = 0; //reads number
  32 long long prevNum = 0;  //not used
  33 int ins_size_var = 20;      // SD of insert-size
  34 PE_INFO * pes = NULL;     //record the pe info in lib file
  35 MEM_MANAGER * rv_mem_manager = NULL;
  36 MEM_MANAGER * cn_mem_manager = NULL;
  37 MEM_MANAGER * arc_mem_manager = NULL;
  38 unsigned int num_vt = 0;                // num of the end-kmer
  39 unsigned long long new_num_vt = 0;      // the new num of the end-kmer after adding the new end-kmer
  40 unsigned int ** found_routes = NULL;
  41 unsigned int * so_far = NULL;       // recorf the path of contig while filling gap
  42 int max_n_routes = 10;
  43 int num_trace;
  44 Kmer WORDFILTER;        //mask code for extracting Kmer info from raw data (two unsigned long long int)
  45 unsigned int num_ed = 0;  //number of edges
  46 unsigned int num_ctg = 0;               // num of contig
  47 unsigned int num_ed_limit;              // the count of edge
  48 unsigned int extraEdgeNum;          // the new count of edge after adding the new edge
  49 EDGE * edge_array = NULL;           // used to record all the info of edge
  50 VERTEX * vt_array = NULL;               // used to record the sequence info of the end-kmer
  51 unsigned int * index_array = NULL;      // used to translate the old contig index to the new contig index
  52 CONTIG * contig_array = NULL;       // used to record all the info of contig
  53 int lineLen;
  54 int len_bar = 100;
  55 int weakPE = 3;     // the minimun weight requirement for the connection
  56 int fillGap = 0;        // 1 for fill the gap after scaffold asm
  57 boolean globalFlag;
  58 long long arcCounter;       // record the num of the arc
  59 MEM_MANAGER * prearc_mem_manager = NULL;
  60 MEM_MANAGER ** preArc_mem_managers = NULL;
  61 int maxReadLen = 0;       //max length will be used for each LIB, soapdenovo read LIBs one by one , for each set a maxReadLen
  62 int maxReadLen4all = 0;   //max length will be used for all reads
  63 int minReadLen = 0;     // min length will be used for all readss
  64 int maxNameLen = 0;       //max length for the name of reads or sequences
  65 ARC ** arcLookupTable = NULL;
  66 long long * markersArray = NULL;
  67 boolean deLowKmer = 0;  //remove the kmers which coverage are not bigger than deLowKmer
  68 boolean deLowEdge = 1;  //remove the edges which coverage are not bigger than deLowEdge
  69 long long newCntCounter;        // record the number of the new connection in one insert-size
  70 long long discardCntCounter;
  71 boolean repsTie = 0;            //sovle tiny repeat or not
  72 CONNECT ** cntLookupTable = NULL;
  73 int num_libs = 0;                   //number of LIBs in read config file
  74 LIB_INFO * lib_array = NULL;        //store LIB's info into lib_array
  75 int libNo = 0;                  // the current number of lib
  76 long long readNumBack;
  77 int gradsCounter;                  //pair number in lib file
  78 unsigned int ctg_short = 0;     //shortest contig for scaffolding
  79 int thrd_num = 8;                  //thread number
  80 int cvgAvg = 0; // the average coverage of contigs
  81 KmerSet ** KmerSets = NULL;      //KmerSet [i] for  thread i
  82 KmerSet ** KmerSetsPatch = NULL; //KmerSet for (k+1) mer
  83 DARRAY * readSeqInGap = NULL;
  84 DARRAY * gapSeqDarray = NULL;
  85 DARRAY ** darrayBuf;
  86 boolean orig2new;   // 1 for re-arrange the contig index using the length
  87 int maxSteps;
  88 boolean maskRep = 1;        // 1 for masking repeat for scaffold asm , 0 for un-masking repeat.
  89 int GLDiff = 50;
  90 int initKmerSetSize = 0;   // init_size = (ubyte8) ((double) initKmerSetSize * 1024.0f * 1024.0f * 1024.0f / (double) thrd_num / 24);
  91 long known_genome_size = 0;
  92 int smallKmer = 0;  // the kmer of the step "Map"
  93 int deltaKmer = 0;  // for map, K-k
  94
  95 double cvg_low = 0.1;
  96 double cvg_high = 2;
  97 double len_times = 0;
  98
  99 float ins_var_idx = 1.5;
 100 int Insert_size = 0;        // the current insert-size
 101 int score_mask = 1;
 102 int COMPATIBLE_MODE = 0;     // 1 for the gz file ; 0 for the normal file
 103 float cvg4SNP = 0.6;
 104
 105 MEM_MANAGER * edgeid_mem_manager = NULL;
 106 unsigned int num_vtnew = 0; //new vertex num
 107 unsigned int kmer_cnew = 0; //new kmer num
 108 const int step = 1;     //step for multi kmer
 109 //int nowstep = 1;
 110 int nowstep2 = 1;
 111
 112 unsigned int * edge_id = NULL;  //edge id array
 113 VERTEX * vt_arraynew = NULL;    //vertex array for k+1mer
 114
 115 KmerSet2 * KmerSetsNew = NULL; //kmer set for k+1mer
 116 char libfilename[256];
 117 boolean parse = 0;
 118 unsigned int num_ed_temp = 0;   //  record the count of the edge
 119
 120 int arcfilter = 0;  //arc filter thrd
 121 boolean outputContig = 0;
 122
 123 long long pinCounter;       //the count of the merged bubble
 124 int clean;  //merge clean bubble
 125
 126 unsigned int num_kmer_limit;
 127
 128 int gLineLen = 5000;
 129 char * gStr = NULL;