additional example ?

[Nick-Mul]

Hi this looks really useful and very easy to use. I am sure this should be obvious but the notebook ends with:

Instead of training a machine learning model, we can also compose the elements to make a pipeline that computes the molecular fingerprint matrix.

this is exactly want I want to do, do you mind give an example of how do output this to file ?

Many thanks,
Nick

[c-w-feldmann]

Hi @Nick-Mul ,

For a simple pipeline returning the Morgan Fingerprint it would look like this:

from molpipeline import Pipeline
from molpipeline.any2mol import AutoToMol
from molpipeline.mol2any import MolToMorganFP

pipeline = Pipeline(
    [
        ("auto2mol", AutoToMol()),
        ("morgan_fp", MolToMorganFP()),
    ]
)

pipeline.fit_transform(your_smiles_list)

The default parameter for MolToMorganFP is return_as=sparse which will return the fingerprints as scipy.sparse.csr_martix. Other options would be dense or explicit_bit_vect, which will return a numpy array or a list of rdkit ExplicitBitVect, respectively.

I hope this answers your Question.

Christian

[Nick-Mul]

Thanks Christian, that's perfect!