Source Code for Module model_selection.overall_disc

  1  # Calculate the overall discrepancy. 
  2  # 
  3  # The input relaxation data for this method should be the operating model data (the theoretical, back calculated 
  4  # relaxation values) which has been Gaussian randomised for a given error.  The original operating model data 
  5  # should be placed in the file 'op_data'.  The format of the file should be as follows.  The first line is a 
  6  # header line and is ignored.  The columns are: 
  7  #       0 - Residue number. 
  8  #       1 - Residue name. 
  9  #       2+ - The real data.  Each column from 2 on should correspond to the data specified in the file 'input'. 
 10  #       When the data in 'input' is 'none', there should be no corresponding column. 
 11  # 
 12  # 
 13  # The program is divided into the following stages: 
 14  #       Stage 1:   Creation of the files for the model-free calculations for models 1 to 5 for the randomised 
 15  #               data. 
 16  #       Stage 2a:  Calculation of the overall discrepancy for model selection, and the creation of the final 
 17  #               run.  Monte Carlo simulations are used to find errors, and the diffusion tensor is unoptimised. 
 18  #               Files are placed in the directory 'final'. 
 19  #       Stage 2b:  Calculation of the overall discrepancy for model selection, and the creation of the final 
 20  #               optimization run.  Monte Carlo simulations are used to find errors, and the diffusion tensor 
 21  #               is optimised.  Files are placed in the directory 'optimised'. 
 22  #       Stage 3:   Extraction of optimised data. 
 23   
 24  import sys 
 25  from math import log, pi 
 26   
 27  from common_ops import common_operations 
 28   
 29   
 30 -class overall_disc(common_operations): 
 31 -        def __init__(self, mf): 
 32                  "Calculation of the theoretical overall discrepancy." 
 33   
 34                  self.mf = mf 
 35   
 36   
 37 -        def model_selection(self): 
 38                  "Model selection." 
 39   
 40                  data = self.mf.data.data 
 41                  self.mf.data.calc_frq() 
 42                  self.mf.data.calc_constants() 
 43                  n = float(self.mf.data.num_data_sets) 
 44                  tm = float(self.mf.usr_param.tm['val']) * 1e-9 
 45   
 46                  if self.mf.debug: 
 47                          self.mf.log.write("\n\n<<< " + self.mf.usr_param.method + " model selection >>>\n\n") 
 48   
 49                  for res in range(len(self.mf.data.relax_data[0])): 
 50                          self.mf.data.results.append({}) 
 51   
 52                          if self.mf.debug: 
 53                                  self.mf.log.write('%-22s\n' % ( "Checking res " + data['m1'][res]['res_num'] )) 
 54   
 55                          real = [] 
 56                          err = [] 
 57                          types = [] 
 58                          for i in range(self.mf.data.num_ri): 
 59                                  real.append(float(self.mf.data.overall_disc.op_data[res][i+2])) 
 60                                  err.append(float(self.mf.data.relax_data[i][res][3])) 
 61                                  types.append([self.mf.data.data_types[i], float(self.mf.data.frq[self.mf.data.remap_table[i]])]) 
 62   
 63                          for model in self.mf.data.runs: 
 64                                  back_calc = [] 
 65                                  for i in range(self.mf.data.num_ri): 
 66                                          label_fit = self.mf.data.frq_label[self.mf.data.remap_table[i]] + "_" + self.mf.data.data_types[i] + "_fit" 
 67                                          back_calc.append(float(self.mf.data.data[model][res][label_fit])) 
 68   
 69                                  chi2 = self.mf.calc_chi2.relax_data(real, err, back_calc) 
 70   
 71                                  if self.mf.debug: 
 72                                          self.mf.log.write("\nReal: " + `real`) 
 73                                          self.mf.log.write("\nError: " + `err`) 
 74                                          self.mf.log.write("\nBack calc: " + `back_calc`) 
 75   
 76                                  data[model][res]['crit'] = chi2 / (2.0 * n) 
 77   
 78                          # Select model. 
 79                          min = 'm1' 
 80                          for run in self.mf.data.runs: 
 81                                  if data[run][res]['crit'] < data[min][res]['crit']: 
 82                                          min = run 
 83                          if data[min][res]['crit'] == float('inf'): 
 84                                  self.mf.data.results[res] = self.fill_results(data[min][res], model='0') 
 85                          else: 
 86                                  self.mf.data.results[res] = self.fill_results(data[min][res], model=min[1]) 
 87   
 88                          if self.mf.debug: 
 89                                  self.mf.log.write(self.mf.usr_param.method + " (m1): " + `data['m1'][res]['crit']` + "\n") 
 90                                  self.mf.log.write(self.mf.usr_param.method + " (m2): " + `data['m2'][res]['crit']` + "\n") 
 91                                  self.mf.log.write(self.mf.usr_param.method + " (m3): " + `data['m3'][res]['crit']` + "\n") 
 92                                  self.mf.log.write(self.mf.usr_param.method + " (m4): " + `data['m4'][res]['crit']` + "\n") 
 93                                  self.mf.log.write(self.mf.usr_param.method + " (m5): " + `data['m5'][res]['crit']` + "\n") 
 94                                  self.mf.log.write("The selected model is: " + min + "\n\n") 
 95   
 96   
 97 -        def print_data(self): 
 98                  "Print all the data into the 'data_all' file." 
 99   
100                  file = open('data_all', 'w') 
101                  file_crit = open('crit', 'w') 
102   
103                  sys.stdout.write("[") 
104                  for res in range(len(self.mf.data.results)): 
105                          sys.stdout.write("-") 
106                          file.write("\n\n<<< Residue " + self.mf.data.results[res]['res_num']) 
107                          file.write(", Model " + self.mf.data.results[res]['model'] + " >>>\n") 
108                          file.write('%-20s' % '') 
109                          file.write('%-19s' % 'Model 1') 
110                          file.write('%-19s' % 'Model 2') 
111                          file.write('%-19s' % 'Model 3') 
112                          file.write('%-19s' % 'Model 4') 
113                          file.write('%-19s' % 'Model 5') 
114   
115                          file_crit.write('%-6s' % self.mf.data.results[res]['res_num']) 
116                          file_crit.write('%-6s' % self.mf.data.results[res]['model']) 
117   
118                          # S2. 
119                          file.write('\n%-20s' % 'S2') 
120                          for run in self.mf.data.runs: 
121                                  file.write('%9.3f' % self.mf.data.data[run][res]['s2']) 
122                                  file.write('%1s' % '±') 
123                                  file.write('%-9.3f' % self.mf.data.data[run][res]['s2_err']) 
124   
125                          # S2f. 
126                          file.write('\n%-20s' % 'S2f') 
127                          for run in self.mf.data.runs: 
128                                  file.write('%9.3f' % self.mf.data.data[run][res]['s2f']) 
129                                  file.write('%1s' % '±') 
130                                  file.write('%-9.3f' % self.mf.data.data[run][res]['s2f_err']) 
131   
132                          # S2s. 
133                          file.write('\n%-20s' % 'S2s') 
134                          for run in self.mf.data.runs: 
135                                  file.write('%9.3f' % self.mf.data.data[run][res]['s2s']) 
136                                  file.write('%1s' % '±') 
137                                  file.write('%-9.3f' % self.mf.data.data[run][res]['s2s_err']) 
138   
139                          # te. 
140                          file.write('\n%-20s' % 'te') 
141                          for run in self.mf.data.runs: 
142                                  file.write('%9.2f' % self.mf.data.data[run][res]['te']) 
143                                  file.write('%1s' % '±') 
144                                  file.write('%-9.2f' % self.mf.data.data[run][res]['te_err']) 
145   
146                          # Rex. 
147                          file.write('\n%-20s' % 'Rex') 
148                          for run in self.mf.data.runs: 
149                                  file.write('%9.3f' % self.mf.data.data[run][res]['rex']) 
150                                  file.write('%1s' % '±') 
151                                  file.write('%-9.3f' % self.mf.data.data[run][res]['rex_err']) 
152   
153                          # Chi2. 
154                          file.write('\n%-20s' % 'Chi2') 
155                          for run in self.mf.data.runs: 
156                                  file.write('%-19.3f' % self.mf.data.data[run][res]['chi2']) 
157   
158                          # Model selection criteria. 
159                          file.write('\n%-20s' % self.mf.usr_param.method) 
160                          for run in self.mf.data.runs: 
161                                  file.write('%-19.6f' % self.mf.data.data[run][res]['crit']) 
162   
163                                  file_crit.write('%-25s' % `self.mf.data.data[run][res]['crit']`) 
164                          file_crit.write('\n') 
165   
166                  file.write('\n') 
167                  sys.stdout.write("]\n") 
168                  file.close() 
169