Post translation modification Protein sequence functional similarity calculator

Question

[ahteshamabbasi1996]

Hi everyone,

I am working with post-translational modification analyses, where I am trying to get the protein sequences (only chunks where PTMS exist) functional similarity scores to form an adjacency matrix for its further use in graph neural networks.

In the past, some works have been related to Mirna where different databases and methods have been used to calculate functional similarities on the basis of weights, etc

If possible, can anyone suggest me to some helpful material as I feel I am lost while doing this kind of analysis?

Another question would be, whether I should use whole protein sequence for such a task, or only the chunk of the sequence where the PTM is present.