Re: [Rdkit-discuss] question about morgan bits

[Greg L. <gre...@gm...>]

Hi Wendong,

The morgan fingerprint algorithm removes redundant atom environments
(environments which contain exactly the same atoms/bonds).
For example, when looking at valine:
[image: image.png]
The environments with radius 2 which are centered on atoms 5 and 6 are
redundant with the environment of radius 1 which is centered on atom 4, so
those environments are not reported in the output.

This is described in more detail in the ECFP paper: Rogers, D.; Hahn, M.
“Extended-Connectivity Fingerprints.” *J. Chem. Inf. and Model.*
**50**:742-54 (2010).

Best,
-greg

On Fri, Mar 12, 2021 at 4:16 AM Wendong Wang <wen...@sj...>
wrote:

> Greetings,
> I have a question about morgan fingerprints. The code is pasted at the end
> of the email, and please see the attached images for the results.
>
> For valine molecule, the radius is set to be 2. The dictionary (atom
> index, radius) shows all the substructures of all atoms with radius 0 as
> fingerprints, and all the substructures of all the atoms with radius 1 as
> fingerprints. But there are only a few substructures with radius 2 as
> fingerprints. Why so few?
>
> Thanks.
>
> Best wishes,
> Wendong
>
> PS. The code is below:
> m1 = Chem.MolFromSmiles('CC(C)[C@@H](C(=O)O)N')
> di1 = {}
> fp1 = AllChem.GetMorganFingerprintAsBitVect(m1, radius = 2, nBits = 2048,
> bitInfo = di1)
> tu1 = [(m1, x, di1) for x in fp1.GetOnBits()]
> Draw.DrawMorganBits(tu1, molsPerRow = 4, legends=[str(x) for x in
> fp1.GetOnBits()])
>
> _______________________________________________
> Rdkit-discuss mailing list
> Rdk...@li...
> https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
>