Generating COSMO-RS Compounds from XYZ Files and SMILES¶

The ADFCOSMORSCompoundJob class generates results equivalent to the GUI task COSMO-RS Compound and provides a convenient Python interface for creating .coskf files for one or many structures.

Downloads: Notebook | Script ?

Requires: AMS2026 or later

Generating coskf files from xyz¶

The example will first load all the molecules in the folder compounds_xyz and then optimize the gas geometry using ADF, and perform the ADF COSMO calculation for each compound. When the calculations are finished, we will find all the .coskf files in the test_coskfs_xyz directory.

from scm.plams import from_smiles, read_molecules, JobRunner, config
from scm.plams.recipes.adfcosmorscompound import ADFCOSMORSCompoundJob
import os
from pathlib import Path

Enable the parallel calculation through JobRunner. Here, we’ll assign one core to each job, and we can have up to eight jobs running all at once.

config.default_jobrunner = JobRunner(
    parallel=True, maxjobs=8
)  # Set the default jobrunner to be parallel
config.default_jobmanager.settings.hashing = None  # Disable rerun prevention
config.job.runscript.nproc = 1  # Number of cores for each job
config.log.stdout = 1  # Suppress plams output

Optional compound metadata keyed by molecule name. This metadata is passed to ADFCOSMORSCompoundJob via mol_info, written to the “Compound Data” section of the generated COSKF file, and used by AMSCrs GUI and pyCRS.Database to search compounds.

mol_info_by_name = {
    "CO": {"CAS": "630-08-0", "IUPAC": "Carbon monoxide", "Other Name": ""},
    "H2O": {"CAS": "7732-18-5", "IUPAC": "Water", "Other Name": ""},
}

compounds_dir = Path("./compounds_xyz")
compounds_dir.mkdir(exist_ok=True)

co_xyz = compounds_dir / "CO.xyz"
h2o_xyz = compounds_dir / "H2O.xyz"

co_xyz.write_text("""\
6

C      -0.36450646      -0.00849588       0.01778348
O       0.93310187      -0.11594431      -0.58640848
H      -1.09583501      -0.23127873      -0.76871971
H      -0.56296738       1.00752571       0.39904705
H      -0.50297820      -0.73401384       0.83725517
H       1.59318518       0.08220705       0.10104249
""")
h2o_xyz.write_text("""\
3

O       0.00615254       0.40173439       0.00000000
H      -0.76825199      -0.18914855       0.00000000
H       0.76209945      -0.21258584       0.00000000
""");

molecules = read_molecules("./compounds_xyz")

results = []
for name, mol in molecules.items():
    mol_info = mol_info_by_name.get(name)
    job = ADFCOSMORSCompoundJob(
        molecule=mol, coskf_name=name, coskf_dir="test_coskfs_xyz", mol_info=mol_info
    )
    results.append(job.run())

[09.04|16:23:22] JOB plamsjob STARTED
[09.04|16:23:22] JOB plamsjob STARTED
[09.04|16:23:22] JOB plamsjob/gas STARTED
[09.04|16:23:22] JOB plamsjob/solv STARTED

for result in results:
    result.wait()

[09.04|16:23:22] JOB plamsjob/sigma STARTED
[09.04|16:23:22] Waiting for job plamsjob to finish
[09.04|16:23:22] JOB plamsjob.002/gas STARTED
[09.04|16:23:22] JOB plamsjob.002/solv STARTED
[09.04|16:23:22] Waiting for job gas to finish
[09.04|16:23:22] JOB plamsjob.002/sigma STARTED
[09.04|16:23:22] Waiting for job solv to finish
[09.04|16:23:22] Waiting for job gas to finish
[09.04|16:23:22] Waiting for job solv to finish
[09.04|16:23:26] JOB plamsjob.002/gas SUCCESSFUL
[09.04|16:23:29] JOB plamsjob.002/solv SUCCESSFUL
[09.04|16:23:29] JOB plamsjob.002/sigma SUCCESSFUL
[09.04|16:23:32] JOB plamsjob.002 SUCCESSFUL
[09.04|16:23:36] JOB plamsjob/gas SUCCESSFUL
[09.04|16:23:48] JOB plamsjob/solv SUCCESSFUL
[09.04|16:23:48] JOB plamsjob/sigma SUCCESSFUL
[09.04|16:23:52] JOB plamsjob SUCCESSFUL

print(f"coskf files generated: {', '.join([f for f in os.listdir('./test_coskfs_xyz')])}")

coskf files generated: CO.coskf, H2O.coskf

Generating .coskf files from smiles¶

Now, we will specify the smiles and name of a set of compounds and generate the initial geometry of each compound using from_smiles function. With the setting, nconfs=100 and forcefield='uff', we will generate 100 conformers and find the one with the lowest energy using ‘uff’ forcefield. When the calculations are finished, we will find all the .coskf file in the test_coskfs_smiles directory.

rd_smiles = ["O", "CO"]
rd_names = ["H2O", "CO"]
molecules = {}
for name, smiles in zip(rd_names, rd_smiles):
    molecules[name] = from_smiles(smiles, nconfs=100, forcefield="uff")[
        0
    ]  # lowest energy one in 100 conformers

Lastly, we give this information to the ADFCOSMORSCompoundJob class, including the name of the coskf files as well as the directory in which we’ll find them after the calculations complete. Using the setting, preoptimization='GFN1-xTB' and singlepoint=False, it will utilize the DFTB for a quick pre-optimization. Subsequently, it will execute a gas phase optimization using ADF, followed by the solvation calculation.

results = []
for name, mol in molecules.items():
    mol_info = mol_info_by_name.get(name)
    job = ADFCOSMORSCompoundJob(
        molecule=mol,  # The initial structure
        coskf_name=name,  # a name to be used for coskf file
        coskf_dir="test_coskfs_smiles",  # a directory to put the .coskf files generated
        preoptimization="GFN1-xTB",  # perform preoptimize or not
        singlepoint=False,  # run a singlepoint in gasphase and solvation calculation without geometry optimization. Cannot be combined with `preoptimization`
        name=name,  # an optional name for the calculation directory
        mol_info=mol_info,  # compound information to be stored in the Compound Data section of the COSKF file
    )
    results.append(job.run())

[09.04|16:23:52] JOB H2O STARTED
[09.04|16:23:52] JOB CO STARTED
[09.04|16:23:52] JOB H2O/preoptimization STARTED
[09.04|16:23:52] JOB H2O/gas STARTED

for result in results:
    result.wait()

[09.04|16:23:52] Waiting for job H2O to finish
[09.04|16:23:52] JOB H2O/solv STARTED
[09.04|16:23:52] JOB H2O/sigma STARTED
[09.04|16:23:52] JOB CO/preoptimization STARTED
[09.04|16:23:52] JOB CO/gas STARTED
[09.04|16:23:52] JOB CO/solv STARTED
[09.04|16:23:52] JOB CO/sigma STARTED
[09.04|16:23:52] Waiting for job gas to finish
[09.04|16:23:52] Waiting for job preoptimization to finish
[09.04|16:23:52] Waiting for job solv to finish
... output trimmed ....
[09.04|16:23:52] JOB CO/preoptimization SUCCESSFUL
[09.04|16:23:59] JOB H2O/gas SUCCESSFUL
[09.04|16:24:04] JOB H2O/solv SUCCESSFUL
[09.04|16:24:04] JOB H2O/sigma SUCCESSFUL
[09.04|16:24:07] JOB H2O SUCCESSFUL
[09.04|16:24:07] Waiting for job CO to finish
[09.04|16:24:21] JOB CO/gas SUCCESSFUL
[09.04|16:24:33] JOB CO/solv SUCCESSFUL
[09.04|16:24:34] JOB CO/sigma SUCCESSFUL
[09.04|16:24:37] JOB CO SUCCESSFUL

print(f"coskf files generated: {', '.join([f for f in os.listdir('./test_coskfs_smiles')])}")

coskf files generated: CO.coskf, H2O.coskf

Python Script¶

#!/usr/bin/env python
# coding: utf-8

# ## Generating coskf files from xyz

# The example will first load all the molecules in the folder ``compounds_xyz`` and then optimize the gas geometry using ADF, and perform the ADF COSMO calculation for each compound. When the calculations are finished, we will find all the .coskf files in the ``test_coskfs_xyz`` directory.

from scm.plams import from_smiles, read_molecules, JobRunner, config
from scm.plams.recipes.adfcosmorscompound import ADFCOSMORSCompoundJob
import os
from pathlib import Path


# Enable the parallel calculation through `JobRunner`. Here, we'll assign one core to each job, and we can have up to eight jobs running all at once.

config.default_jobrunner = JobRunner(parallel=True, maxjobs=8)  # Set the default jobrunner to be parallel
config.default_jobmanager.settings.hashing = None  # Disable rerun prevention
config.job.runscript.nproc = 1  # Number of cores for each job
config.log.stdout = 1  # Suppress plams output


# Optional compound metadata keyed by molecule name. This metadata is passed to ADFCOSMORSCompoundJob via `mol_info`, written to the "Compound Data" section of the generated COSKF file, and used by AMSCrs GUI and `pyCRS.Database` to search compounds.

mol_info_by_name = {
    "CO": {"CAS": "630-08-0", "IUPAC": "Carbon monoxide", "Other Name": ""},
    "H2O": {"CAS": "7732-18-5", "IUPAC": "Water", "Other Name": ""},
}


compounds_dir = Path("./compounds_xyz")
compounds_dir.mkdir(exist_ok=True)

co_xyz = compounds_dir / "CO.xyz"
h2o_xyz = compounds_dir / "H2O.xyz"

co_xyz.write_text("""\
6

C      -0.36450646      -0.00849588       0.01778348
O       0.93310187      -0.11594431      -0.58640848
H      -1.09583501      -0.23127873      -0.76871971
H      -0.56296738       1.00752571       0.39904705
H      -0.50297820      -0.73401384       0.83725517
H       1.59318518       0.08220705       0.10104249
""")
h2o_xyz.write_text("""\
3

O       0.00615254       0.40173439       0.00000000
H      -0.76825199      -0.18914855       0.00000000
H       0.76209945      -0.21258584       0.00000000
""")
molecules = read_molecules("./compounds_xyz")

results = []
for name, mol in molecules.items():
    mol_info = mol_info_by_name.get(name)
    job = ADFCOSMORSCompoundJob(molecule=mol, coskf_name=name, coskf_dir="test_coskfs_xyz", mol_info=mol_info)
    results.append(job.run())


for result in results:
    result.wait()


print(f"coskf files generated: {', '.join([f for f in os.listdir('./test_coskfs_xyz')])}")


# ## Generating .coskf files from smiles

# Now, we will specify the smiles and name of a set of compounds and generate the initial geometry of each compound using `from_smiles` function. With the setting, `nconfs=100` and `forcefield='uff'`, we will generate 100 conformers and find the one with the lowest energy using 'uff' forcefield. When the calculations are finished, we will find all the .coskf file in the ``test_coskfs_smiles`` directory.

rd_smiles = ["O", "CO"]
rd_names = ["H2O", "CO"]
molecules = {}
for name, smiles in zip(rd_names, rd_smiles):
    molecules[name] = from_smiles(smiles, nconfs=100, forcefield="uff")[0]  # lowest energy one in 100 conformers


# Lastly, we give this information to the `ADFCOSMORSCompoundJob` class, including the name of the coskf files as well as the directory in which we'll find them after the calculations complete.  Using the setting, `preoptimization='GFN1-xTB'` and `singlepoint=False`, it will utilize the DFTB for a quick pre-optimization. Subsequently, it will execute a gas phase optimization using ADF, followed by the solvation calculation.

results = []
for name, mol in molecules.items():
    mol_info = mol_info_by_name.get(name)
    job = ADFCOSMORSCompoundJob(
        molecule=mol,  # The initial structure
        coskf_name=name,  # a name to be used for coskf file
        coskf_dir="test_coskfs_smiles",  # a directory to put the .coskf files generated
        preoptimization="GFN1-xTB",  # perform preoptimize or not
        singlepoint=False,  # run a singlepoint in gasphase and solvation calculation without geometry optimization. Cannot be combined with `preoptimization`
        name=name,  # an optional name for the calculation directory
        mol_info=mol_info,  # compound information to be stored in the Compound Data section of the COSKF file
    )
    results.append(job.run())


for result in results:
    result.wait()


print(f"coskf files generated: {', '.join([f for f in os.listdir('./test_coskfs_smiles')])}")

Generating COSMO-RS Compounds from XYZ Files and SMILES¶

Generating coskf files from xyz¶

Generating .coskf files from smiles¶

See also¶

Related examples¶

Related tutorials¶

Related documentation¶

Python Script¶