• Interpreting your results
• Tendencies
• File formats
• FAQ
• Related papers

Interpreting your results from Interface Affinity: 


The IncreaseAffinity protocol is based on the premise that increasing buried hydrophobic surface area and/or decreasing 
buried hydrophilic surface area will generally lead to enhanced binding affinity provided steric clashes are not introduced 
and buried polar groups are not left without a hydrogen bond partner. All point mutations at the protein-protein interface
that fit the above criteria are tried, and the resulting changes in protein-protein binding free energies (ddGbinding) are
computed.  ddGbinding  are computed using the following equation:
ddGbinding=dGmut-dGwt  where dGmut/dGwt are the binding energies computed for the mutant and wild type structures respectively.   
An energies table is output by Rosetta.  The results in the table are sorted, with the point mutations that are predicted to yield
the greatest increase in binding affinity at the top of the table.  The results are also filtered such that point mutations that
are predicted to give a ddGbinding of greater than -0.5 kcal/mol are not included.

Increase Affinity tendencies:


This protocol assumes that the conformation of the unbound state is the same as the bound state, 
thus losing information about the stability of individual chains in the ddGbinding predictions.  
We therefore include ddGchain A and ddGchain B in the output table.  We have also found that the 
energetic penalty for removing a hydrogen bond in an unminimized crystal structure can be insignificant 
compared to the other energy terms, resulting in a false prediction that the mutation is stabilizing 
to the complex.  In addition, if a polar residue has a large number of neighbors (~>20), that residue 
can be considered very buried, and is likely participating in a hydrogen bond and should probably not be 
mutated. We include the ddGh-bond and number of neighbors to help the user screen for these types of 
destabilizing mutations. 
The following additional filters may be applied at the user's discretion, to address the above mentioned 
issues in an automated way: 1) require the ddGh-bond, the hydrogen bond energy term, to be zero or less, 
directing the results to exclude mutations that remove a side-chain involved in a hydrogen bond, and 
2) require that the ddGchain A, and ddGchain B be less than or equal to 1 in an effort to remove mutations 
that are predicted to significantly destabilize one of the individual chains.
Our work indicates that adding these filters does tend to increase the rate of success, but at the expense 
of the number of predictions.  Some protein-protein interfaces yield so few predictions that pass these filters 
that it becomes necessary to leave one or both of them off.

INTOUT file from IncreaseAffinity:


Mutation  ddG_bind  ddG_chainA  ddG_chainB  #Neighbors  ddG_h_bond
 98B N>L:    -9.7       0.0        0.3         14            0.0
 98B N>Y:    -9.7       0.0        0.3         14            0.0
 98B N>F:    -9.6       0.0        0.4         14            0.0
 98B N>M:    -9.4       0.0        0.6         14            0.0

The mutation data is in following format:    
98B N>L where the chain is B, the wild type amino acid is N, the sequence position is 98, 
and the mutant amino acid is L.

FAQs:
1.Could RosettaDesign calculate native protein free energy only instead of doing a design?
Yes, RosettaDesign can calculate native free energy. The easiest thing to do is to submit a job which doesn't vary the sequence
( using "create list" option and make all residues fixed ).

2. What is MAX_RES in the log file?
MAX_RES in the log file means that the total number of residues in your native protein. The current limitation of MAX_RES is 1000.

3. What is the difference between "MAX_RES" and "maximum residues can be varied"?
"MAX_RES" is the total number of residues in your native protein. The current limitation of MAX_RES is 1000.
"maximum residues can be varied" is the number of residues you choose to be redesigned.

4. What does "can't find starting residue in pdb file" mean in the log file?
"can't find starting residue in pdb file" usually means that your uploaded file's format is not a typical pdb file so that the server can
not find the start residue to redesign.

5. What is "unrecognized residue:"?
The server read the three letters representation of 20 types of amino acids identities from the pdb file. Sometimes, there are one or more
identities which are not included in the 20 amino acides shows in the pdb files so that the server can not recognized them.

6. What is "missing backbone atoms"?
"missing backbone atoms" means there are part of back bone atoms are missed in the pdb file so that the server can not get correct information
to redesign the backbone.

Papers:
1: Design of a novel globular protein fold with atomic-level accuracy.
Kuhlman B, et al.
Science. 2003 Nov 21;302(5649):1364-8.
PMID: 14631033 [PubMed - in process]
 
2: A large scale test of computational protein design: folding and stability of nine completely redesigned globular proteins.
Dantas G, et al.
J Mol Biol. 2003 Sep 12;332(2):449-60.
PMID: 12948494 [PubMed - indexed for MEDLINE]
 
3: Crystal structures and increased stabilization of the protein G variants with switched folding pathways NuG1 and NuG2.
Nauli S, et al.
Protein Sci. 2002 Dec;11(12):2924-31.
PMID: 12441390 [PubMed - indexed for MEDLINE]
 
4: Accurate computer-based design of a new backbone conformation in the second turn of protein L.
Kuhlman B, et al.
J Mol Biol. 2002 Jan 18;315(3):471-7.
PMID: 11786026 [PubMed - indexed for MEDLINE]
 
5: Conversion of monomeric protein L to an obligate dimer by computational protein design.
Kuhlman B, et al.
Proc Natl Acad Sci U S A. 2001 Sep 11;98(19):10687-91. Epub 2001 Aug 28. Erratum in: Proc Natl Acad Sci U S A 2002 May 28;99(11):7809.
PMID: 11526208 [PubMed - indexed for MEDL INE]
 
6: Computer-based redesign of a protein folding pathway.
Nauli S, et al.
Nat Struct Biol. 2001 Jul;8(7):602-5.
PMID: 11427890 [PubMed - indexed for MEDLINE]
 
7: Native protein sequences are close to optimal for their structures.
Kuhlman B, et al.
Proc Natl Acad Sci U S A. 2000 Sep 12;97(19):10383-8. Erratum in: Proc Natl Acad Sci U S A. 2000 Nov 21;97(24):13460.
PMID: 10984534 [PubMed - indexed for MEDLINE]