Preprocessing of amino acid sequences

[TrikiSadok]

I am not that familiar with RDKit.
Does RDKIt allow transforming amino acid sequences into numerical features for machine learning purposes?
example of amino acid sequence I am trying to transform into numerical descriptor that is included in the data set I have:
"
MIGIDIVSIARVEKCVKRFEMRFLERFLSPSEIVLCKDKSSSIAGFFALKEACSKALQVGIGKELSFLDMRISKSPKNAPLITLSKEKMDYFNIQSLSASISHDAGFAIAVVMISP
"
Thanks in advance for any help.

[greglandrum]

It does! You can convert a sequence into an RDKit molecule as follows:

>>> m = Chem.MolFromSequence('MIGIDIVSIARVEKCVKRFEMRFLERFLSPSEIVL')

And then all the standard RDKit descriptors can be used:

>>> from rdkit.Chem import rdMolDescriptors
>>> rdMolDescriptors.CalcNumAmideBonds(m)
34

[TrikiSadok]

Thanks a lot.
Is there a difference in treatment between MolFromSequence() and MolFromFASTA() in this case ?? since I found in other forums people are using MolFromFASTA() to transform amino acid and protein sequences.

[greglandrum]

No. Other than the obvious fact that a different parser is used for the different formats, the molecules are processed the same way.

[TrikiSadok]

very clear, Thanks a lot.