binaries/src/clustalo/src/hhalign/hhmatrices-C.h

   1 /* -*- mode: c; tab-width: 4; c-basic-offset: 4; indent-tabs-mode: nil -*- */
   2
   3 /*********************************************************************
   4  * Clustal Omega - Multiple sequence alignment
   5  *
   6  * Copyright (C) 2010 University College Dublin
   7  *
   8  * Clustal-Omega is free software; you can redistribute it and/or
   9  * modify it under the terms of the GNU General Public License as
  10  * published by the Free Software Foundation; either version 2 of the
  11  * License, or (at your option) any later version.
  12  *
  13  * This file is part of Clustal-Omega.
  14  *
  15  ********************************************************************/
  16
  17 /*
  18  * RCS $Id: hhmatrices-C.h 199 2011-02-21 18:24:49Z fabian $
  19  */
  20
  21 // Substitution matrices and their background frequencies
  22
  23 // The following background frequencies were calculated by the formula pa = (P[a,b]/(pa*pb))^(-1) * (1,...,1)
  24 // For the Blousum50-matrix this becomes  pb[a]= SUM_(b=1,20) (2^(BLOSUM50[a,b]/3))^(-1)
  25 //                     A    R    N    D    C    Q    E    G    H    I    L     K    M    F    P    S    T    W    Y    V
  26 // Gonnet            7.68,5.14,4.02,5.41,1.89,3.27,5.99,7.56,3.69,5.06,10.01,5.97,2.20,3.50,4.54,4.67,7.12,1.25,3.95,7.28
  27 // BLOSUM50          8.24,6.24,4.46,4.77,2.03,2.90,6.78,6.69,2.53,6.89,10.7 ,5.04,1.49,4.93,3.97,5.95,6.13,1.34,3.45,6.28
  28
  29 const float Gonnet[]={
  30 //  A     R     N     D     C     Q     E     G     H     I     L     K     M     F     P     S     T     W     Y     V
  31  10227, 3430, 2875, 3869, 1625, 2393, 4590, 6500, 2352, 3225, 5819, 4172, 1435, 1579, 3728, 4610, 6264,  418, 1824, 5709, // A
  32   3430, 7780, 2209, 2589,  584, 2369, 3368, 3080, 2173, 1493, 3093, 5701,  763,  859, 1893, 2287, 3487,  444, 1338, 2356, // R
  33   2875, 2209, 3868, 3601,  501, 1541, 2956, 3325, 1951, 1065, 2012, 2879,  532,  688, 1480, 2304, 3204,  219, 1148, 1759, // N
  34   3869, 2589, 3601, 8618,  488, 2172, 6021, 4176, 2184, 1139, 2151, 3616,  595,  670, 2086, 2828, 3843,  204, 1119, 2015, // D
  35   1625,  584,  501,  488, 5034,  355,  566,  900,  516,  741, 1336,  591,  337,  549,  419,  901, 1197,  187,  664, 1373, // C
  36   2393, 2369, 1541, 2172,  355, 1987, 2891, 1959, 1587, 1066, 2260, 2751,  570,  628, 1415, 1595, 2323,  219,  871, 1682, // Q
  37   4590, 3368, 2956, 6021,  566, 2891, 8201, 3758, 2418, 1624, 3140, 4704,  830,  852, 2418, 2923, 4159,  278, 1268, 2809, // E
  38   6500, 3080, 3325, 4176,  900, 1959, 3758,26066, 2016, 1354, 2741, 3496,  741,  797, 2369, 3863, 4169,  375, 1186, 2569, // G
  39   2352, 2173, 1951, 2184,  516, 1587, 2418, 2016, 5409, 1123, 2380, 2524,  600, 1259, 1298, 1642, 2446,  383,  876, 1691, // H
  40   3225, 1493, 1065, 1139,  741, 1066, 1624, 1354, 1123, 6417, 9630, 1858, 1975, 2225, 1260, 1558, 3131,  417, 1697, 7504, // I
  41   5819, 3093, 2012, 2151, 1336, 2260, 3140, 2741, 2380, 9630,25113, 3677, 4187, 5540, 2670, 2876, 5272, 1063, 3945,11005, // L
  42   4172, 5701, 2879, 3616,  591, 2751, 4704, 3496, 2524, 1858, 3677, 7430,  949,  975, 2355, 2847, 4340,  333, 1451, 2932, // K
  43   1435,  763,  532,  595,  337,  570,  830,  741,  600, 1975, 4187,  949, 1300, 1111,  573,  743, 1361,  218,  828, 2310, // M
  44   1579,  859,  688,  670,  549,  628,  852,  797, 1259, 2225, 5540,  975, 1111, 6126,  661,  856, 1498, 1000, 4464, 2602, // F
  45   3728, 1893, 1480, 2086,  419, 1415, 2418, 2369, 1298, 1260, 2670, 2355,  573,  661,11834, 2320, 3300,  179,  876, 2179, // P
  46   4610, 2287, 2304, 2828,  901, 1595, 2923, 3863, 1642, 1558, 2876, 2847,  743,  856, 2320, 3611, 4686,  272, 1188, 2695, // S
  47   6264, 3487, 3204, 3843, 1197, 2323, 4159, 4169, 2446, 3131, 5272, 4340, 1361, 1498, 3300, 4686, 8995,  397, 1812, 5172, // T
  48    418,  444,  219,  204,  187,  219,  278,  375,  383,  417, 1063,  333,  218, 1000,  179,  272,  397, 4101, 1266,  499, // W
  49   1824, 1338, 1148, 1119,  664,  871, 1268, 1186,  876, 1697, 3945, 1451,  828, 4464,  876, 1188, 1812, 1266, 9380, 2227, // Y
  50   5709, 2356, 1759, 2015, 1373, 1682, 2809, 2569, 1691, 7504,11005, 2932, 2310, 2602, 2179, 2695, 5172,  499, 2227,11569};// V
  51
  52 const float Blosum30[]= {
  53 //    A      R      N      D      C      Q      E      G      H      I      L      K      M      F      P      S      T      W      Y      V
  54   0.0096,
  55   0.0038,0.0109,
  56   0.0031,0.0019,0.0055,
  57   0.0043,0.0026,0.0027,0.0095,
  58   0.0014,0.0011,0.0010,0.0010,0.0070,
  59   0.0031,0.0031,0.0014,0.0018,0.0007,0.0039,
  60   0.0044,0.0031,0.0024,0.0037,0.0018,0.0028,0.0094,
  61   0.0052,0.0032,0.0032,0.0035,0.0011,0.0020,0.0035,0.0173,
  62   0.0016,0.0014,0.0011,0.0011,0.0004,0.0010,0.0018,0.0015,0.0060,
  63   0.0040,0.0022,0.0024,0.0018,0.0012,0.0016,0.0023,0.0036,0.0012,0.0072,
  64   0.0056,0.0039,0.0032,0.0043,0.0023,0.0027,0.0051,0.0051,0.0022,0.0066,0.0139,
  65   0.0044,0.0043,0.0027,0.0032,0.0010,0.0021,0.0053,0.0039,0.0013,0.0026,0.0044,0.0063,
  66   0.0018,0.0012,0.0009,0.0008,0.0004,0.0007,0.0012,0.0013,0.0008,0.0014,0.0027,0.0017,0.0012,
  67   0.0027,0.0023,0.0017,0.0011,0.0008,0.0010,0.0016,0.0022,0.0008,0.0027,0.0055,0.0023,0.0008,0.0077,
  68   0.0028,0.0021,0.0012,0.0021,0.0007,0.0015,0.0030,0.0030,0.0014,0.0017,0.0027,0.0029,0.0005,0.0011,0.0091,
  69   0.0056,0.0035,0.0028,0.0034,0.0013,0.0021,0.0038,0.0051,0.0017,0.0033,0.0047,0.0042,0.0010,0.0027,0.0024,0.0075,
  70   0.0040,0.0019,0.0024,0.0024,0.0009,0.0016,0.0023,0.0028,0.0010,0.0027,0.0046,0.0025,0.0010,0.0016,0.0020,0.0041,0.0046,
  71   0.0005,0.0007,0.0002,0.0004,0.0003,0.0004,0.0007,0.0011,0.0002,0.0005,0.0009,0.0006,0.0002,0.0007,0.0004,0.0005,0.0003,0.0027,
  72   0.0014,0.0022,0.0008,0.0015,0.0004,0.0011,0.0016,0.0016,0.0010,0.0018,0.0045,0.0017,0.0007,0.0024,0.0011,0.0017,0.0014,0.0009,0.0044,
  73   0.0056,0.0031,0.0022,0.0027,0.0012,0.0015,0.0026,0.0033,0.0012,0.0063,0.0074,0.0030,0.0015,0.0032,0.0017,0.0036,0.0036,0.0005,0.0024,0.0083};
  74
  75 const float Blosum40[]= {
  76 //    A      R      N      D      C      Q      E      G      H      I      L      K      M      F      P      S      T      W      Y      V
  77   0.0148,
  78   0.0029,0.0109,
  79   0.0029,0.0019,0.0069,
  80   0.0032,0.0021,0.0031,0.0126,
  81   0.0015,0.0007,0.0007,0.0009,0.0093,
  82   0.0026,0.0024,0.0017,0.0016,0.0004,0.0048,
  83   0.0040,0.0026,0.0023,0.0048,0.0010,0.0032,0.0118,
  84   0.0066,0.0023,0.0035,0.0031,0.0012,0.0020,0.0028,0.0260,
  85   0.0014,0.0012,0.0013,0.0014,0.0003,0.0010,0.0015,0.0014,0.0060,
  86   0.0037,0.0017,0.0017,0.0016,0.0008,0.0012,0.0018,0.0024,0.0009,0.0105,
  87   0.0050,0.0030,0.0021,0.0027,0.0015,0.0023,0.0036,0.0034,0.0017,0.0082,0.0209,
  88   0.0041,0.0051,0.0027,0.0030,0.0010,0.0026,0.0045,0.0035,0.0013,0.0023,0.0037,0.0099,
  89   0.0017,0.0010,0.0007,0.0007,0.0003,0.0007,0.0010,0.0013,0.0007,0.0018,0.0037,0.0012,0.0018,
  90   0.0022,0.0016,0.0013,0.0013,0.0008,0.0008,0.0015,0.0021,0.0009,0.0031,0.0055,0.0018,0.0011,0.0105,
  91   0.0027,0.0014,0.0014,0.0019,0.0005,0.0013,0.0026,0.0028,0.0008,0.0020,0.0021,0.0023,0.0007,0.0010,0.0151,
  92   0.0060,0.0025,0.0030,0.0030,0.0013,0.0026,0.0034,0.0052,0.0013,0.0025,0.0034,0.0034,0.0011,0.0019,0.0022,0.0089,
  93   0.0040,0.0019,0.0022,0.0024,0.0011,0.0015,0.0025,0.0029,0.0009,0.0028,0.0039,0.0029,0.0011,0.0019,0.0022,0.0043,0.0070,
  94   0.0007,0.0005,0.0003,0.0003,0.0001,0.0004,0.0005,0.0008,0.0002,0.0005,0.0009,0.0005,0.0002,0.0008,0.0003,0.0004,0.0003,0.0045,
  95   0.0019,0.0016,0.0010,0.0011,0.0004,0.0010,0.0014,0.0017,0.0012,0.0020,0.0031,0.0017,0.0010,0.0032,0.0009,0.0015,0.0014,0.0008,0.0060,
  96   0.0054,0.0023,0.0017,0.0022,0.0012,0.0014,0.0025,0.0028,0.0008,0.0083,0.0082,0.0027,0.0018,0.0031,0.0018,0.0032,0.0040,0.0005,0.0020,0.0113};
  97
  98 const float Blosum50[]= {
  99 //    A      R      N      D      C      Q      E      G      H      I      L      K      M      F      P      S      T      W      Y      V
 100   0.0192,
 101   0.0027,0.0152,
 102   0.0024,0.0020,0.0101,
 103   0.0026,0.0019,0.0035,0.0161,
 104   0.0015,0.0005,0.0006,0.0005,0.0091,
 105   0.0022,0.0025,0.0016,0.0017,0.0004,0.0057,
 106   0.0034,0.0029,0.0023,0.0048,0.0006,0.0033,0.0141,
 107   0.0062,0.0020,0.0031,0.0028,0.0009,0.0017,0.0023,0.0316,
 108   0.0012,0.0013,0.0015,0.0011,0.0003,0.0010,0.0013,0.0011,0.0064,
 109   0.0035,0.0015,0.0013,0.0012,0.0008,0.0011,0.0015,0.0018,0.0007,0.0140,
 110   0.0048,0.0028,0.0017,0.0018,0.0014,0.0019,0.0026,0.0027,0.0013,0.0104,0.0304,
 111   0.0033,0.0064,0.0027,0.0026,0.0006,0.0029,0.0043,0.0028,0.0014,0.0017,0.0027,0.0130,
 112   0.0016,0.0009,0.0007,0.0005,0.0004,0.0008,0.0008,0.0009,0.0005,0.0022,0.0042,0.0010,0.0029,
 113   0.0020,0.0012,0.0009,0.0008,0.0006,0.0007,0.0012,0.0015,0.0009,0.0030,0.0058,0.0012,0.0012,0.0154,
 114   0.0022,0.0011,0.0011,0.0015,0.0004,0.0011,0.0019,0.0019,0.0006,0.0013,0.0017,0.0018,0.0005,0.0007,0.0171,
 115   0.0062,0.0025,0.0032,0.0028,0.0011,0.0022,0.0030,0.0044,0.0012,0.0021,0.0029,0.0031,0.0010,0.0016,0.0018,0.0111,
 116   0.0039,0.0021,0.0026,0.0022,0.0010,0.0015,0.0024,0.0025,0.0009,0.0029,0.0038,0.0026,0.0011,0.0015,0.0016,0.0047,0.0100,
 117   0.0005,0.0004,0.0002,0.0002,0.0001,0.0003,0.0004,0.0005,0.0002,0.0005,0.0008,0.0004,0.0003,0.0009,0.0002,0.0003,0.0004,0.0059,
 118   0.0015,0.0013,0.0009,0.0009,0.0004,0.0009,0.0012,0.0012,0.0013,0.0018,0.0027,0.0013,0.0007,0.0039,0.0006,0.0013,0.0012,0.0008,0.0077,
 119   0.0054,0.0020,0.0015,0.0016,0.0013,0.0014,0.0020,0.0022,0.0007,0.0107,0.0092,0.0022,0.0021,0.0030,0.0016,0.0029,0.0041,0.0005,0.0018,0.0164};
 120
 121 const float Blosum65[]= {
 122 //  A      R      N      D      C      Q      E      G      H      I      L      K      M      F      P      S      T      W      Y      V
 123   0.0222,
 124   0.0022,0.0181,
 125   0.0019,0.0019,0.0148,
 126   0.0021,0.0015,0.0037,0.0225,
 127   0.0016,0.0004,0.0004,0.0004,0.0127,
 128   0.0018,0.0024,0.0015,0.0016,0.0003,0.0076,
 129   0.0029,0.0025,0.0021,0.0049,0.0003,0.0034,0.0168,
 130   0.0057,0.0016,0.0027,0.0024,0.0007,0.0013,0.0018,0.0396,
 131   0.0010,0.0013,0.0014,0.0009,0.0002,0.0010,0.0013,0.0009,0.0096,
 132   0.0031,0.0012,0.0009,0.0011,0.0012,0.0008,0.0012,0.0013,0.0006,0.0191,
 133   0.0043,0.0023,0.0013,0.0014,0.0016,0.0016,0.0019,0.0020,0.0009,0.0115,0.0388,
 134   0.0032,0.0062,0.0024,0.0024,0.0005,0.0030,0.0040,0.0024,0.0012,0.0015,0.0024,0.0166,
 135   0.0013,0.0007,0.0005,0.0004,0.0004,0.0007,0.0006,0.0007,0.0003,0.0025,0.0052,0.0008,0.0045,
 136   0.0016,0.0009,0.0007,0.0007,0.0005,0.0005,0.0008,0.0011,0.0008,0.0030,0.0056,0.0009,0.0012,0.0186,
 137   0.0021,0.0009,0.0008,0.0011,0.0003,0.0008,0.0014,0.0013,0.0005,0.0009,0.0013,0.0015,0.0004,0.0005,0.0195,
 138   0.0065,0.0022,0.0030,0.0027,0.0011,0.0018,0.0029,0.0037,0.0011,0.0017,0.0024,0.0030,0.0008,0.0012,0.0016,0.0137,
 139   0.0037,0.0017,0.0022,0.0019,0.0009,0.0013,0.0021,0.0022,0.0007,0.0027,0.0033,0.0023,0.0010,0.0012,0.0013,0.0048,0.0133,
 140   0.0004,0.0003,0.0002,0.0001,0.0002,0.0002,0.0002,0.0004,0.0002,0.0004,0.0007,0.0003,0.0002,0.0009,0.0001,0.0003,0.0003,0.0074,
 141   0.0013,0.0009,0.0007,0.0006,0.0004,0.0006,0.0008,0.0008,0.0016,0.0015,0.0023,0.0010,0.0006,0.0043,0.0004,0.0010,0.0009,0.0010,0.0113,
 142   0.0049,0.0015,0.0011,0.0012,0.0014,0.0011,0.0016,0.0017,0.0006,0.0120,0.0094,0.0019,0.0023,0.0025,0.0012,0.0023,0.0035,0.0004,0.0015,0.0206};
 143
 144
 145 const float Blosum80[]= {
 146 //  A      R      N      D      C      Q      E      G      H      I      L      K      M      F      P      S      T      W      Y      V
 147   0.0252,
 148   0.0020,0.0210,
 149   0.0016,0.0017,0.0166,
 150   0.0018,0.0013,0.0037,0.0262,
 151   0.0015,0.0003,0.0004,0.0003,0.0172,
 152   0.0017,0.0024,0.0014,0.0014,0.0003,0.0094,
 153   0.0028,0.0023,0.0019,0.0048,0.0003,0.0035,0.0208,
 154   0.0053,0.0015,0.0025,0.0022,0.0006,0.0011,0.0017,0.0463,
 155   0.0009,0.0012,0.0012,0.0008,0.0002,0.0011,0.0012,0.0008,0.0104,
 156   0.0027,0.0010,0.0007,0.0008,0.0011,0.0007,0.0010,0.0009,0.0004,0.0220,
 157   0.0036,0.0018,0.0011,0.0011,0.0014,0.0014,0.0015,0.0016,0.0008,0.0111,0.0442,
 158   0.0029,0.0061,0.0022,0.0020,0.0004,0.0028,0.0036,0.0020,0.0010,0.0012,0.0019,0.0190,
 159   0.0011,0.0006,0.0004,0.0003,0.0004,0.0007,0.0006,0.0005,0.0003,0.0025,0.0052,0.0007,0.0053,
 160   0.0014,0.0007,0.0006,0.0006,0.0005,0.0005,0.0006,0.0009,0.0007,0.0027,0.0052,0.0007,0.0010,0.0211,
 161   0.0021,0.0009,0.0007,0.0009,0.0003,0.0007,0.0012,0.0010,0.0004,0.0007,0.0012,0.0012,0.0003,0.0004,0.0221,
 162   0.0064,0.0020,0.0029,0.0024,0.0010,0.0017,0.0026,0.0034,0.0010,0.0015,0.0021,0.0026,0.0007,0.0010,0.0014,0.0167,
 163   0.0036,0.0015,0.0020,0.0016,0.0009,0.0012,0.0019,0.0019,0.0007,0.0024,0.0028,0.0020,0.0009,0.0011,0.0011,0.0048,0.0156,
 164   0.0003,0.0002,0.0001,0.0001,0.0001,0.0002,0.0002,0.0003,0.0001,0.0003,0.0006,0.0002,0.0002,0.0007,0.0001,0.0002,0.0002,0.0087,
 165   0.0011,0.0007,0.0006,0.0005,0.0003,0.0005,0.0006,0.0006,0.0016,0.0013,0.0020,0.0008,0.0005,0.0046,0.0003,0.0009,0.0008,0.0010,0.0148,
 166   0.0046,0.0013,0.0009,0.0010,0.0013,0.0010,0.0015,0.0014,0.0005,0.0123,0.0089,0.0015,0.0022,0.0022,0.0010,0.0021,0.0033,0.0004,0.0012,0.0246};
 167
 168
 169 // prediction accuracy of Psipred:
 170 // Ppred[cf][B][A] = P(A,B,cf)/P(A)/P(B,cf) = P(A|B,cf)/P(A)
 171 // A = observed ss state  B = predicted ss state  cf = confidence value of prediction
 172 //float Ppred[MAXCF][NSSPRED][NDSSP]=
 173 const float Ppred[]=
 174   {
 175 //pred/obs  -      H      E      ~      S      T      G      B
 176         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=-
 177         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // H
 178         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // E
 179         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // ~
 180         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=0
 181         1.000, 1.128, 0.519, 0.834, 0.957, 1.488, 2.106, 1.085 , // H
 182         1.000, 0.233, 2.240, 1.216, 0.913, 0.519, 0.923, 1.759 , // E
 183         1.000, 0.640, 1.017, 1.122, 1.069, 1.242, 2.140, 1.999 , // ~
 184         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=1
 185         1.000, 1.251, 0.485, 0.771, 0.847, 1.371, 2.266, 0.864 , // H
 186         1.000, 0.222, 2.542, 1.069, 0.804, 0.428, 0.671, 1.728 , // E
 187         1.000, 0.474, 1.103, 1.295, 1.232, 1.214, 1.835, 1.989 , // ~
 188         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=2
 189         1.000, 1.383, 0.426, 0.637, 0.778, 1.349, 2.436, 0.824 , // H
 190         1.000, 0.202, 2.769, 0.999, 0.714, 0.320, 0.551, 1.566 , // E
 191         1.000, 0.395, 1.005, 1.407, 1.376, 1.336, 1.725, 2.063 , // ~
 192         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=3
 193         1.000, 1.531, 0.369, 0.552, 0.682, 1.280, 2.420, 0.698 , // H
 194         1.000, 0.169, 2.970, 0.954, 0.556, 0.273, 0.489, 1.504 , // E
 195         1.000, 0.352, 0.843, 1.515, 1.542, 1.456, 1.684, 1.958 , // ~
 196         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=4
 197         1.000, 1.750, 0.305, 0.444, 0.537, 1.134, 2.295, 0.600 , // H
 198         1.000, 0.124, 3.179, 0.847, 0.513, 0.228, 0.413, 1.897 , // E
 199         1.000, 0.282, 0.718, 1.664, 1.630, 1.577, 1.625, 1.877 , // ~
 200         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=5
 201         1.000, 1.952, 0.250, 0.353, 0.456, 0.982, 2.050, 0.466 , // H
 202         1.000, 0.102, 3.464, 0.699, 0.453, 0.174, 0.284, 1.357 , // E
 203         1.000, 0.227, 0.574, 1.782, 1.846, 1.681, 1.418, 1.885 , // ~
 204         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=6
 205         1.000, 2.183, 0.171, 0.263, 0.319, 0.792, 1.933, 0.345 , // H
 206         1.000, 0.079, 3.712, 0.612, 0.281, 0.133, 0.196, 1.089 , // E
 207         1.000, 0.173, 0.458, 1.915, 2.007, 1.766, 1.220, 1.704 , // ~
 208         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=7
 209         1.000, 2.389, 0.132, 0.192, 0.224, 0.605, 1.605, 0.183 , // H
 210         1.000, 0.053, 3.997, 0.449, 0.201, 0.072, 0.141, 0.919 , // E
 211         1.000, 0.109, 0.328, 2.013, 2.304, 1.882, 0.960, 1.512 , // ~
 212         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=8
 213         1.000, 2.668, 0.065, 0.098, 0.144, 0.354, 1.059, 0.102 , // H
 214         1.000, 0.029, 4.285, 0.284, 0.113, 0.044, 0.059, 0.522 , // E
 215         1.000, 0.053, 0.200, 2.099, 2.444, 2.133, 0.671, 1.290 , // ~
 216         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=9
 217         1.000, 2.966, 0.009, 0.023, 0.036, 0.113, 0.214, 0.017 , // H
 218         1.000, 0.010, 4.555, 0.119, 0.027, 0.010, 0.013, 0.209 , // E
 219         1.000, 0.026, 0.101, 2.576, 1.853, 2.204, 0.308, 0.499   // ~
 220   };
 221
 222 // float Ppred[]=
 223 //   {
 224 // //pred/obs  -      H      E      ~      S      T      G      B
 225 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=-
 226 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 227 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 228 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 229 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=0
 230 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 231 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 232 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 233 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=1
 234 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 235 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 236 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 237 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=2
 238 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 239 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 240 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 241 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=3
 242 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 243 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 244 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 245 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=4
 246 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 247 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 248 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 249 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=5
 250 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 251 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 252 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 253 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=6
 254 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 255 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 256 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 257 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=7
 258 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 259 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 260 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 261 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=8
 262 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 263 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 264 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 265 //         1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000 , // - conf=9
 266 //         1.000, 2.000, 0.500, 1.000, 1.000, 1.000, 1.000, 0.500,  // H
 267 //         1.000, 0.500, 2.000, 0.500, 1.000, 1.000, 0.500, 1.000,  // E
 268 //         1.000, 1.000, 1.000, 2.000, 2.000, 2.000, 1.000, 1.000,  // ~
 269
 270
 271 float Pobs[]={0, 0.3268,0.2119,0.2061,0.0913,0.1143,0.0376,0.0120};
 272
 273 float Pevo_full[]=
 274   {
 275 //     -     H     E     C
 276     1.00, 0.00, 0.00, 0.00,
 277     0.00, 0.94, 0.00, 0.04,
 278     0.00, 0.00, 0.92, 0.04,
 279     0.00, 0.06, 0.08, 0.92
 280   };
 281
 282 //psipred accuracy for confidence values 0-9
 283 const float p_acc[]={0.00,0.47,0.53,0.56,0.58,0.62,0.69,0.74,0.82,0.88,0.96};
 284
 285 /**
 286  * @brief
 287  */
 288 void
 289 SetBlosumMatrix(const float BlosumXX[])
 290 {
 291   int a,b,n=0;
 292   if (v>=3) printf("Using the BLOSUM%2i matrix\n",par.matrix);
 293   for (a=0; a<20; ++a)
 294     for (pb[a]=0.0f, b=0; b<=a; ++b,++n)
 295       P[a][b] = BlosumXX[n];
 296   for (a=0; a<19; a++)
 297     for (b=a+1; b<20; ++b)
 298       P[a][b] = P[b][a];
 299   for (a=0; a<20; ++a) P[a][20]=P[20][a]=1.0f;
 300   return;
 301 }
 302
 303 /////////////////////////////////////////////////////////////////////////////////////
 304 /**
 305  * @brief Set (global variable) substitution matrix with derived matrices and background frequencies
 306  */
 307 void
 308 SetSubstitutionMatrix()
 309 {
 310   int a,b;
 311   switch (par.matrix)
 312     {
 313     default:
 314     case 0:  //Gonnet matrix
 315       if (v>=3) cout<<"Using the Gonnet matrix ";
 316       for (a=0; a<20; ++a)
 317         for (pb[a]=0.0f, b=0; b<20; ++b)
 318           P[a][b] = 0.000001f*Gonnet[a*20+b];
 319       for (a=0; a<20; ++a) P[a][20]=P[20][a]=1.0f;
 320       break;
 321
 322     case 30:  //BLOSUM30
 323       SetBlosumMatrix(Blosum30);
 324       break;
 325     case 40:  //BLOSUM40
 326       SetBlosumMatrix(Blosum40);
 327       break;
 328     case 50:  //BLOSUM50
 329       SetBlosumMatrix(Blosum50);
 330       break;
 331     case 65:  //BLOSUM65
 332       SetBlosumMatrix(Blosum65);
 333       break;
 334     case 80:  //BLOSUM80
 335       SetBlosumMatrix(Blosum80);
 336       break;
 337    }
 338
 339   // Check transition probability matrix, renormalize P and calculate pb[a]
 340   float sumab=0.0f;
 341   for (a=0; a<20; a++)
 342     for (b=0; b<20; ++b) sumab+=P[a][b];
 343   for (a=0; a<20; a++)
 344     for (b=0; b<20; ++b) P[a][b]/=sumab;
 345   for (a=0; a<20; a++)
 346     for (pb[a]=0.0f, b=0; b<20; ++b) pb[a]+=P[a][b];
 347
 348   //Compute similarity matrix for amino acid pairs (for calculating consensus sequence)
 349   for (a=0; a<20; ++a)
 350     for (b=0; b<20; ++b)
 351       Sim[a][b] = P[a][b]*P[a][b]/P[a][a]/P[b][b];
 352
 353   //Precompute matrix R for amino acid pseudocounts:
 354   for (a=0; a<20; ++a)
 355     for (b=0; b<20; ++b)
 356       R[a][b] = P[a][b]/pb[b]; //R[a][b]=P(a|b)
 357
 358   //Precompute matrix R for amino acid pseudocounts:
 359   for (a=0; a<20; ++a)
 360     for (b=0; b<20; ++b)
 361       S[a][b] = log2(R[a][b]/pb[a]); // S[a][b] = log2(P(a,b)/P(a)/P(b))
 362
 363   // Evaluate sequence identity underlying substitution matrix
 364   if (v>=3)
 365     {
 366       float id=0.0f;
 367       float entropy=0.0f;
 368       float entropy_pb=0.0f;
 369       float mut_info=0.0f;
 370       for (a=0; a<20; ++a) id+=P[a][a];
 371       for (a=0; a<20; ++a) entropy_pb-=pb[a]*log2(pb[a]);
 372       for (a=0; a<20; ++a)
 373           for (b=0; b<20; ++b)
 374             {
 375               entropy-=P[a][b]*log2(R[a][b]);
 376               mut_info += P[a][b]*S[a][b];
 377             }
 378
 379       printf(": sequence identity = %2.0f%%; entropy per column = %4.2f bits (out of %4.2f); mutual information = %4.2f bits\n",100*id,entropy,entropy_pb,mut_info);
 380     }
 381
 382   if (v>=4) //Debugging: probability matrix and dissimilarity matrix
 383     {
 384       cout<<"Check matrix: before renormalization sum P(a,b)= "<<sumab<<"...\n";//PRINT
 385       cout<<"      A    R    N    D    C    Q    E    G    H    I    L    K    M    F    P    S    T    W    Y    V\n";
 386       cout<<"p[] ";
 387       for (a=0; a<20; a++)  printf("%4.1f ",100*pb[a]);
 388       cout<<endl<<"\nSubstitution matrix log2( P(a,b)/p(a)/p(b) ) (in bits):\n";
 389       cout<<"      A    R    N    D    C    Q    E    G    H    I    L    K    M    F    P    S    T    W    Y    V\n";
 390       for (b=0; b<20; b++)
 391         {
 392           cout<<i2aa(b)<<"   ";
 393           for (a=0; a<20; a++)  printf("%4.1f ",S[a][b]);
 394           cout<<endl;
 395         }
 396       cout<<endl<<"\nOdds matrix P(a,b)/p(a)/p(b):\n";
 397       cout<<"      A    R    N    D    C    Q    E    G    H    I    L    K    M    F    P    S    T    W    Y    V\n";
 398       for (b=0; b<20; b++)
 399         {
 400           cout<<i2aa(b)<<"   ";
 401           for (a=0; a<20; a++)  printf("%4.1f ",P[b][a]/pb[a]/pb[b]);
 402           cout<<endl;
 403         }
 404       cout<<endl<<"\nMatrix of conditional probabilities P(a|b) = P(a,b)/p(b) (in %):\n";
 405       cout<<"      A    R    N    D    C    Q    E    G    H    I    L    K    M    F    P    S    T    W    Y    V\n";
 406       for (b=0; b<20; b++)
 407         {
 408           cout<<i2aa(b)<<"   ";
 409           for (a=0; a<20; a++)  printf("%4.1f ",100*R[b][a]);
 410           cout<<endl;
 411         }
 412       cout<<endl<<"\nProbability matrix P(a,b) (in %):\n";
 413       cout<<"      A     R     N     D     C     Q     E     G     H     I     L     K     M     F     P     S     T     W     Y     V\n";
 414       for (b=0; b<20; b++)
 415         {
 416           cout<<i2aa(b)<<"   ";
 417           for (a=0; a<20; a++)  printf("%5.0f ",1000000*P[b][a]);
 418           cout<<endl;
 419         }
 420       cout<<endl<<"Similarity matrix P(a,b)^2/P(a,a,)/P(b,b) (in %):\n";
 421       cout<<"      A    R    N    D    C    Q    E    G    H    I    L    K    M    F    P    S    T    W    Y    V\n";
 422       for (b=0; b<20; b++)
 423         {
 424           cout<<i2aa(b)<<"   ";
 425           for (a=0; a<20; a++)  printf("%4.0f ",100*Sim[b][a]);
 426           cout<<endl;
 427         }
 428       cout<<endl;
 429
 430
 431     }
 432 }
 433
 434
 435 /////////////////////////////////////////////////////////////////////////////////////
 436 /**
 437  * @brief Set secondary structure substitution matrix
 438  */
 439 void
 440 SetSecStrucSubstitutionMatrix()
 441 {
 442   int A;        //observed ss state (determined dssp)
 443   int B,BB;     //predicted ss states (by psipred)
 444   int cf,ccf;   //confidence value of prediction
 445   float P73[NDSSP][NSSPRED][MAXCF];  //P73[cf][B][A] = P(A,B,cf)/P(A)/P(B,cf) = P(A|B,cf)/P(A)
 446   float sum;
 447
 448   // S73[A][B][cf][b] = score for matching observed ss state A in query with state B in template
 449   // predicted with confidence cf, when query and template columns are diverged by b units
 450   for (cf=0; cf<MAXCF; cf++)
 451     for (A=0; A<NDSSP; A++)
 452       for (B=0; B<NSSPRED; B++)
 453         {
 454           P73[A][B][cf] = 1.-par.ssa + par.ssa*Ppred[cf*NSSPRED*NDSSP + B*NDSSP + A];
 455           S73[A][B][cf] = log2(P73[A][B][cf]);
 456         }
 457
 458   for (B=0; B<NSSPRED; B++)
 459     for (cf=0; cf<MAXCF; cf++)
 460       for (BB=0; BB<NSSPRED; BB++)
 461         for (ccf=0; ccf<MAXCF; ccf++)
 462           {
 463             sum=0;
 464             for (A=1; A<NDSSP; A++)
 465               sum += P73[A][B][cf] * P73[A][BB][ccf] * Pobs[A];
 466             S33[B][cf][BB][ccf] = log2(sum);
 467           }
 468 } /* this is the end of SetSecStrucSubstitutionMatrix() */
 469
 470
 471
 472 /*
 473  * EOF hhmatrices-C.h
 474  */