Changelog

Changelog#

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

unreleased#

3.7.3 - 2024.07.05#

Changed#

  • Name from MolDrug to moldrug (mainly for documentation).

3.7.2 - 2024.01.22#

Fix#

  • It was possible to generate the same children in the same generations. Now there is a filter to avoid repeated evaluations.

3.7.1 - 2024.01.14#

Fix#

  • Remove an unnecessary print statement.

3.7.0 - 2024.01.14#

Fixed#

  • Calculate SA_score with the molecule without explicit hydrogens for all built-in fitness functions. Based on: rdkit/rdkit#7047

Added#

  • The moldrug.constraintconf.clashes_present function.

Changed#

  • bio.PDB object is no longer used for filtering of conformations that clash with the protein. Now the coordinates are retrieved from the RDKit molecule object and the distance is calculated with NumPy.

Removed#

  • bio dependency for constraint conformer generation.

3.6.1 - 2023.12.11#

Fix#

  • LICENCE on PyPi.

Add#

  • Extra metadata to the package.

3.6.0 - 2023.12.07#

Added#

  • vina_seed argument for:

    • all the cost functions of moldrug.fitness,

    • moldrug.fitness._vinadock.

  • randomseed:

    • moldrug.utils.confgen,

    • moldrug.utils.Individual,

    • moldrug.utils.Local,

    • moldrug.utils.GA,

    • moldrug.constraintconf.constraintconf as well its CLI with the flag --seed.

  • -V and --verbose flag in the CLI of moldrug.

  • moldrug.utils.softmax function.

  • moldrug.utils.deep_update function.

  • moldrug.utils.__get_default_desirability` to store the default desirability values.

  • vina_score_type = 'ensemble', this is used by CostMultiReceptors* functions. It is meant to be used to account for flexibility in the receptor. This is equivalent to performing ensemble docking.

Removed#

  • outdir flag in moldrug CLI.

Changed#

  • Use random.random instead numpy.random.rand function for some small cases.

  • AllChem.MMFFOptimizeMolecule is only used internally for moldrug.utils.confgen and moldrug.constraintconf.gen_aligned_conf if randomseed is not set and with maxIters = 500.

  • pickle module is replaced by dill. It is better to handle users’ custom fitness functions.

  • Data now is retrieved with moldrug``.data.get_data function.

  • Is not needed to input the whole desirability definition if only is intended to change one part of the default desirability used internally by the cost functions of moldrug.fitness. For example, if you would like to change the Target value of vina_score from its default value.

    Before:

    ga = GA(
  ...
  costfunc = moldrug.fitness.Cost
  costfunc_kwargs = {
    ...
    "desirability": {
      "qed": {
          "w": 1,
          "LargerTheBest": {
            "LowerLimit": 0.1,
            "Target": 0.75,
            "r": 1
          }
      },
      "sa_score": {
          "w": 1,
          "SmallerTheBest": {
            "Target": 3,
            "UpperLimit": 7,
            "r": 1
          }
      },
      "vina_score": {
          "w": 1,
          "SmallerTheBest": {
            "Target": -12,
            "UpperLimit": -6,
            "r": 1
          }
      }
    },
    ...
  }
  ...
)

    Now:

    ga = GA(
  ...
  costfunc = moldrug.fitness.Cost
  costfunc_kwargs = {
    ...
    "desirability": {
      "vina_score": {
          "SmallerTheBest": {
            "Target": -12,
          }
      }
    },
    ...
  }
  ...
)

    The same from the CLI.

Fixed#

  • Small bug when the population has Individuals with the same cost. Better reproducibility.

  • Refactored changes.

  • Handled properly in case receptor_pdbqt_path = None.

  • Convert self.crem_db_path to absolute path in moldrug.utils.Local and moldrug.utils.GA.

3.5.0 - 2023.10.26#

Fixed#

  • Refactored changes for for meeko-0.5.0.

  • Update docs.

Added#

  • A new environmental variable MOLDRUG_VERBOSE can be set to ask for more debug information.

Changed#

  • From csf to toml package configuration.

  • Structure of the repo. Now src/moldrug.

  • Improve moldrug-Dashboard.

Removed#

  • Support for conda package. Some dependencies had not ben installed properly. Fix in the future.

3.4.0 - 2023.03.10#

Added#

  • Continuation option to the command line.

  • moldrug.cli.CommandLineHelper class to work with the parameters passed through the command line.

  • checkpoint option to moldrug.utils.GA.

  • moldrug-Dashboard add-on. This is not included on the package itself, but could be used online or locally. In the case of locally you must check Streamlit, the requirements.txt and the app script.

  • retunr_mol option to utils.to_dataframe

Changed#

  • The warnings are not printed for: moldrug.fitness._vinadock and moldrug.constraintconf.generate_conformers. Now, only at the end of a moldrug simulation, a note will be printed if the error.tar.gz file is created.

  • moldrug.utils.run does not print extra info if the command fails. It only raises the corresponding RuntimeError.

  • moldrug.fitness.__vinadock by moldrug.fitness._vinadock.

  • Remove conformers that clash with the protein in case of score_only, for local_only vina will handle the clash.

Fixed#

  • Small bug during the initialization of the population with multiple seed_mol. Now seed_mol with the same amount of elements as popsize are not submitted to mutations, only evaluation.

  • Problems with default parameters definition on the command line. Parameters with default values by the type of run defined in the configuration file do not need to be redefined anymore; moldrug.cli will guess those.

3.3.0 - 2022.12.21#

Changed#

  • moldrug.utils.__tar_errors__ by moldrug.utils.tar_errors.

  • The default value of moldrug.utils.tar_errors is error instead of .error.

  • moldrug.constraintconf.generate_conformers outputs warnings and errors to error instead of .error.

  • moldrug.fitness._vinadock outputs warnings and errors to error instead of .error.

Fixed#

  • The use of moldrug.utils.tar_errors inside of moldrug.utils.Local and moldrug.utils.GA.

  • Clean code.

3.2.5 - 2022.12.20#

Fixed#

  • Improve docs.

  • Cleaning of temporal files in /tmp directory. Now it is created temporal directories in the working directory with the pattern: .costfunc_moldrug_XXXXXXXX.

  • Cleaning errors. Now all warnings and errors are saved in .error directory and at the end they are compressed to `error.tar.gz``.

3.2.2 - 2022.12.12#

Fixed#

  • Bug: The initial individual was not printed properly.

Removed#

  • Redundant code in moldrug.utils.GA

3.2.0 - 2022.11.28#

Fixed#

  • Bug: The output error name when constraint fails has an idx prefix. E.g. 33_conf_27_error.pbz2 now is: idx_33_conf_27_error.pbz2. Now it is easy to delete all of these files at the end of the simulation if they are not needed. (on the last version the naming was not changing)

Removed#

  • moldrug.fitness.is_inside_box

  • Bug: constraint_check_inside_box option for the cost functions of moldrug.fitness

3.1.0 - 2022.11.28#

Added#

  • moldrug.constraintconf.gen_aligned_conf

  • In case that moldrug.constraintconf.generate_conformers fails with rdkit.Chem.AllChem.ConstrainedEmbedwill try withmoldrug.constraintconf.gen_aligned_conf`.

  • moldrug.fitness.is_inside_box

  • constraint_check_inside_box arguments to the cost functions of moldrug.fitness. If the coordinates of the constraint conformation are outside the box; use always local_only, by default False

Changed#

  • The output error name when constraint fails has an idx prefix. E.g. 33_conf_27_error.pbz2 now is: idx_33_conf_27_error.pbz2. Now it is easy to delete all of these files at the end of the simulation if they are not needed.

Fixed#

  • Clean code.

  • Improve docs.

3.0.3 - 2022.11.26#

Added#

  • Warning in case moldrug.utils.Local or moldrug.utils.GA are called with a different moldrug as they were initialized.

Changed#

  • Convert to absolute path receptor_pdbqt_path and vina_executable (in case that it points to a file) inside of moldruf.fitness.__vinadock.

Fixed#

  • Bug for hydrogen coordinates when constrain docking was used.

  • Improve docs.

3.0.1 - 2022.11.24#

Fixed#

  • Cleaning code.

  • Sort the initial population based on the cost attribute when it is saved on disk.

  • Improve docs.

3.0.0 - 2022.11.23#

Changed#

  • Name of moldrug.fitness.get_mol_cost to moldrug.fitness.__get_mol_cost function.

  • The class moldrug.utils.GA does not have any more the method roulette_wheel_selection anymore; now is part of a function that could be called from moldrug.utils

  • max for max_conf in moldrug.constraintconf.constraintconf() function.

  • Entrance point constraintconf was changed to constraintconf_moldrug and now it is linked to moldrug.cli.__constraintconf_cmd instead moldrug.constrainconf.constraintconf_cmd.

  • Name of the function moldrug.fitness.vinadock now is moldrug.fitness.__vinadock.

  • Name of the function moldrug.cli.moldrug_cmd now is moldrug.cli.__moldrug_cmd.

Fixed#

  • Cleaning the code.

  • If vina_executable is provided (to any cost function) it represents a path. It will try to convert to an absolute path. Previously relative paths to the executable were not understood properly.

  • Improve docs.

Added#

  • ad4map in all the cost functions of the moldrug.fitness module. This parameter specifies the path where the ad4 map files are. To use this feature you must have the AutoDcok Vina v1.2.3 of above. Now you can use the force fields of AD4 inside of Vina. Future releases will extend the integration with these versions.

  • moldrug.utils.to_dataframe. This function was previously isolated as a method of the class moldrug.utils.GA; now it could also be called as a function.

  • kept_gens attribute to the Individuals inside of moldrug.utils.GA. This is a set that contains the generations for which the Individual was conserved.

  • acceptance attribute to moldrug.utils.GA. This is a dictionary that has as a keyword the generation ID, and as values a dictionary with keywords: accepted (number of generated molecules accepted on the current generation) and generated (number of total molecules generated)

  • Print Accepted rate= accepted /  generated during running.

  • Add hydrogens before creating pdbt file with meeko when constrain docking i used.

  • seed_mol of moldrug.utils.GA now could be a list (or iterable in a general way) of RDKit molecules. This feature could be used to combine several moldrug runs and create a final run with this combined population.

  • seed_mol from the command line could be a valid SMILES, a list of valid SMILES or a list of paths to the _pop.pbz2 binary files. In the last case, all the populations will be combined and sorted based on the cost attribute. If the result population is less that popsize new structures will be generated to complete the initial population. The individuals of this initial population will be reinitialized and the cost function will be calculated.

2.1.12 - 2022.09.29#

Added#

  • score_only bool parameter to moldrug.fitness.get_mol_cost. Print the starting date when moldrug is called from the command line.

Removed#

  • Type Hints int for attribute idx on moldrug.utils.Individual.

Changed#

  • If AutoDock-Vina fails inside moldrug.fitness.vinadock; give as pdbqt the string “VinaFailed”.

  • If the molecule has a molecular weight higher than wt_cutoff when moldrug.fitness.CostOnlyVina (or moldrug.fitness.CostMultiReceptorsOnlyVina) is called; the pdbqt attribute of the returned Individual will be the string “TooHeavy” (or the list of strings List[“TooHeavy”])

2.1.7 - 2022.09.02#

Fixed#

  • Bug on moldrug.fitness.vinadock during searching of MCS between Individual.mol and constraint_ref. Before was needed to manually specify the atom IDs of the seed_mol that match constraint_ref, now it is not needed anymore.

  • Bug during handling exception in moldrug.constraintconf.generate_conformers.

  • Bug during handling exception moldrug.constraintconf.generate_conformers in moldrug.fitness.vinadock.

  • Bug(s) when constraint docking is used on different Vina versions. The output of vina is not the same and therefore moldrug.fitness.vinadock failed.

Changed#

  • In case constraint = True in moldrug.fitness.vinadock, ref_smi will be the MCF between individual.mol and constraint_ref instead of the SMILES string of constraint_ref when moldrug.constrainconf.generate_conformers is internally called.

Added#

  • moldrug.fitness.get_mol_cost function.

  • Attribute genID to the generated individuals during a moldrug.utils.GA run.

2.1.0 - 2022.08.30#

Fixed#

  • Bug during the calculation of probabilities when costs are larger numbers.

  • Expose hidden error if some Exception occurred during the parallel run.

Added#

  • moldrug.constrainconf module.

  • Raise ValueError if ref_smi is invalid in moldrug.utils.constrainconf.generate_conformers.

Changed#

  • In case constraint = True in moldrug.fitness.vinadock, ref_smi will be the SMILES string of constraint_ref when moldrug.constrainconf.generate_conformers is internally called. This is in order to avoid errors when moldrug.utils.constrainconf.generate_conformerstries to guessref_smi` based on MCS and fails, see this RDKit bug. The workaround for constraint docking is explained here: Constraint Docking.

  • moldrug.fitness.generate_conformers does not fail. In case of Exception, it returns the same mol without conformers and writes the error in a log file into the working directory.

  • The attribute name bestcost by best_cost of moldrug.utils.GA.

  • The functions duplicate_conformers, get_mcs, generate_conformers, constraintconf and constraintconf_cmd and the class ProteinLigandClashFilter were moved from moldrug.fitness module to moldrug.constrainconf module.

  • Entrance point constraintconf now it is linked to moldrug.constrainconf.constraintconf_cmd instead moldrug.fitness.constraintconf_cmd.

2.0.0 - 2022.08.25#

Added#

  • The functions duplicate_conformers, get_mcs, generate_conformers, constraintconf and constraintconf_cmd and the class ProteinLigandClashFilter. The code was borrowed from Pat Walters. It is used if constraint docking is needed.

  • constraintconf can be called from the command line.

  • moldrug.fitness.vinadock() a simple wrapper around vina. This function will be used for all the implemented cost functions inside of the module moldrug.fitness. It could be used for constraint docking.

  • moldrug.data.constraintref. This module is used for testing in case constraint docking is needed. It has two MolBlock strings: r_6lu7 and r_x0161. That could be easily converted into RDKit molecules.

    from rdkit import Chem
from moldrug.data import constraintref
mol = Chem.MolFromMolBlock(constraintref.r_x0161)

    This molecule is needed for the keyword argument constraint_ref of the functions of the moldrug.fitness module in case of constraint docking is used.

  • Constraint docking capability in all implemented cost functions of the module moldrug.fitness.

  • moldrug.data.receptor_pdb. This module is similar to moldrug.data.receptor_pdbqt but in pdb format.

  • Documentation and tutorials.

Changed#

  • moldrug.utils.make_sdf only will create the sdf file based on the pdbqt attribute. If pdbqt is a list, it will work as the previous version works with pdbqts attribute.

  • Name of the module moldrug.data.receptors to moldrug.data.receptor_pdbqt.

  • Name of keyword argument receptor_path to receptor_pdbqt_path on the cost functions: moldrug.fitness.Cost and moldrug.fitness.CostOnlyVina.

  • Name of keyword arguments receptor_path, vina_score_types, boxcenters and boxsizes to receptor_pdbqt_path, vina_score_type, boxcenter and boxsize respectively on the cost functions: moldrug.fitness.CostMultiReceptors and moldrug.fitness.CostMultiReceptorsOnlyVina.

  • smiles attribute in moldrug.utils.Individual now it is always without explicit Hs, despite if the mol attribute has them.

1.1.0 - 2022.08.23#

Changed#

  • moldrug.utils.Individual now is a hashable object.

  • moldrug.utils.GA.SawIndividuals now is a set instead of a list

  • moldrug.utils.update_reactant_zone sets the keywords matchValences and ringMatchesRingOnly to True on rdFMCS.FindMCS. This prevents undesired effects. E.g:

    from moldrug.utils import update_reactant_zone
from rdkit import Chem
mol1 = Chem.MolFromSmiles('c1ccccc1')
mol2 = Chem.MolFromSmiles('CCCN')
update_reactant_zone(parent=mol1,offspring=mol2 parent_replace_ids = [0,2])

    Before the results was: ([0, 2, 3], []). Now it is: ([0, 1, 2, 3], []). The behavior was because the ringMatchesRingOnly is set to False by default inside RDKit.

Removed#

  • moldrug.utils.timeit. No longer needed.

Added#

  • Documentation.

1.0.2 - 2022.08.08#

Removed#

  • Popen option in moldrug.utils.run.

Changed#

  • RuntimeError by warnings.warn when vina run fails and save every error as idx_error.pbz2. Where idx is the index of the failed individual.

  • Print format when mutate fails inside of moldrug.utils.GA

Added#

  • Print moldrug’s version when the command line is used.

1.0.0 - 2022.07.30#

Fixed#

  • Hidden RuntimeError in moldrug.fitness module.

  • Bug printing number of generations in final info.

Removed#

  • Unused code in moldrug.home module.

  • 3D conformation in the mol attribute of the Individual during initialization.

  • The use of grow_mol in the initialization of the populating when get_similar = True. Now the population is initialized with mutate_mol and the same set of crem parameters used during the searching.

  • The automatic addition of Hs in the case where min_size and/or max_size were equal to zero. Now if your intention is to work with the hydrogens, you must provide a SMILES with the explicit Hs. In the future, the input could be just an RDKit mol. Now you must specify if you would like to add explicit Hs to the molecule with the keyword AddHs; the default is False and is used for both moldrug.utils.GA and moldrug.utils.Local.

Added#

  • Handling vina RuntimeError and keeping track of debug. This feature is used to identify what is the error. In the future will be removed.

  • Two new fitness functions: moldrug.fitness.CostOnlyVina and moldrug.fitness.CostMultiReceptorsOnlyVina. They only use the information of Vina scoring function. See the docs for more info about it.

  • Tracking of atom indexes during generations in order to use protected_ids and replace_ids options of mutate_mol function of CReM. Before it was not possible; the use of these features generated undesired solutions because the indexes are not static over generations. Even so, there are still some problems for symmetric molecules. We are working on it.

Changed#

  • The whole moldrug works base on the RDKit mol instead of the SMILES string:

    1. moldrug.utils.Individual is now initialized mol instead smiles. Now the SMILES string is generated internally, it still used as identifying for the instance.

    2. moldrug.utils.GA changed smiles for seed_mol and it is not needed the mol variable any more.

    3. moldrug.utils.Local changed mol for seed_mol in the initialization variables.

    4. moldrug.utils.confgen changed smiles for mol variable.

0.1.0 - 2022-07-25#

Fixed#

  • Minor code cleaning.

  • Bug during the import of user custom cost function.

Added#

  • outdir option for the command line.

  • User custom desirability.

0.0.4 - 2022-07-21#

Fixed#

  • Minor compatibility issue with Python 3.8 (issue #4).

  • Problem with the user’s custom cost function supplied on the command line.

  • Local class compatible with the command line.

  • Minor code cleaning.

  • Better code covered during testing

Contents