Hello.
My question is a little bit specific so I need to explain the process before asking.
Given an active site on a protein, we want to copy this site on a small protein, in order to create it in our lab. In a larger view, we try to imitate a protein function on a small protein.
For this, I am working on a program which search through the PDB to find small 3D patterns of protein structures. This search is solely based on C-alpha and C-beta distances comparisons. It means you can find similar patterns on the atom level but not at the amino acid level.
At the end, the program gives a list of proteins with a site similar to the reference pattern. These proteins are called "mime" while the original pattern is called "ref".
The problem is the following : When you experimentally mutate an amino acid of a mime to copy the ref pattern, it can lead to a destabilization of the structure of the protein.
To avoid this kind of problems, we filter our list of mime to detect which can be mutated and which cannot.
Now my question !
To test our filtering process, we need a benchmark of couple ref-mime for which we know if mutations in the mime are allowed or not. Do you know where I can find such informations ? some databases, like Protherm ?
It seems to fulfill my wishes, I will test this tool during the week. Many thanks !