# Copyright Mirela Andronescu 2008 # Sample options file for CGlearn.pl # Comments are lines starting with # ###################################################################### # Most important input options ###################################################################### Executable that creates the training and test data sets: tools/add_initial_predictions_simfold # The name of the executable that can be used to obtain the training and test data sets # none if no such executable is provided, in which case training set (and optionally test set) # have to be provided. Training file: data/TRA_StructFull-max500-fold1.txt # Other examples: data/TRA_151Rfam_Tco_MFE_fm363.txt # must contain predictions with the initial parameter set Training filename: TRA-SF1 # Other examples: TRA-151Rfam # the "Training filename" is a friendly short name used for the new directory name only Number of iterations {50|100|...}: 2 # the number of CG iterations Bounds parameter B {1K|10K|0.8%|inf|...}: 1K # lower and upper bound on the parameters # if you don't want any bounds, type inf (for infinity) Thermo file: data/all_thermodynamic_constraints_fm363_l2norm.lp # The file with the thermodynamic constraints Weight of the thermodynamic set lambda {0|1|0.9|0.995}: 0.995 # it is in [0,1]. It is 0 if no thermodynamic set is included # the recommended value is 0.995 Test file: data/TES_StructFull-max500-fold1.txt # the file I test on at the end of the iterations # must be in the same format as the training file # can be none Initial params file: data/turner_parameters_fm363_constrdangles.txt # The file with the initial parameters Fixparams file: data/params_fix_205_259.txt # A file which contains fixed values for some parameters, or the word # "variable" for the non-fixed parameters. # For example data/params_fix_fitted_thermo_l2norm.txt contains values # obtained by doing linear regression on the thermodynamic set only. # Another example is when we want to fix the dangling end parameters # to the Turner99 values or to 0. # type none if no parameters are fixed Fixparams filename: FIX-205-259 # a friendly name used for the new directory name only Additional constraints file: data/constraints_dangling_ends_fm363.txt # A file which contains general constraints. # For example, in the 363-parameters model, we want the 3' dangling ends # to be less than the 5' dangling ends. # Specific bounds on the parameters can be given here too. # for example the dangling ends have to be negative # Example of such file: data/constraints_dangling_ends_fm363.txt # type none if no additional constraints are given Additional constraints filename: CON-DL # A friendly name used for the new directory name only Executable that creates the structural constraints: tools/create_structural_constraints_simfold # The name of the executable that creates the structural constraints Executable that predicts and analyses results: tools/predict_and_analyse_results_simfold # The name of the executable that predicts and analyses the results of the new parameters ###################################################################### # Other options, I suggest you leave these unchanged ###################################################################### L norm used for the objective function {1|2}: 2 # default is 2 Normalisation criterion {n1|n2}: n2 # n1: 1/D, n2 = 1/N*1/numstr (in the ISMB paper it's n2) Model is simplified {1|0}: 0 # 1 if we don't consider internal loops 1x1, 1x2 and 2x2 separately # 0 if we consider the model described in the ISMB paper (363 parameters) Do perturbation {0|1}: 0 # 0 means the algorithm stops if no more constraints can be generated # 1 means the latest parameters are randomly perturbed by at most 1kcal/mol, # and then the iterations continue.