QSAR models

We can call the available QSAR models for a chemical and an end point position selected from the end point tree. Here's how we can get a list of the available QSAR models for estimating the Henry's law constant, and then call the models for a given chemID:

henry_qsar_models = qs.get_qsar_models(position=henry_position)
print(f"Available QSAR models: {henry_qsar_models}")

HLC_QSAR_list = []
for model in henry_qsar_models:
    HLC_QSAR = {}
    res = qs.apply_qsar_chemical(
        res_chem_cas[0]["ChemId"], qsar_model_guid=model["Guid"]
    )
    HLC_QSAR[model["Caption"]] = res["Value"]
    HLC_QSAR["Unit"] = res["Unit"]
    HLC_QSAR_list.append(HLC_QSAR)
print(HLC_QSAR_list)

The above values are estimated with the available QSAR methods for the Henry's law constant in the QSAR Toolbox. However, there is another way to do it, which is using calculators.
* Note: the OPERA interface takes longer to run (probably for converting SMILES to structure), between 2-5 minutes. Therefore, I suggest not running it from this interface and running it directly from OPERA CLI. We will release a Python package for doing it automatically.

calculators_list = qs.get_available_calculators()
henry_calculators = []
henry_calculator_values = []
for c in calculators_list:
    if "henry" in c["Caption"].lower():
        h = {}
        h["Caption"] = c["Caption"]
        res = qs.run_calculator(
            calculation_guid=c["Guid"], chem_id=res_chem_cas[0]["ChemId"]
        )
        h["Value"] = res["Value"]
        h["Unit"] = res["Unit"]
        henry_calculator_values.append(h)
        henry_calculators.append(c)
print(henry_calculator_values)

Although not systematically tested, the calculators seem to run faster than models, and therefore can be recommended for longer list of chemicals.

Next step

An step by step guide for obtaining experimental data and QSAR model results from QSAR Toolbox via Python.