RosettaDesign (each run produces the same sequence with many GLY residues )

[Anonymous]

Hi All,

I have some questions about rosetta design. I have been trying to use rosettaDesign “fixbb” to design a cyclic-peptide that binds to an enzyme. Here are the questions I have:

1) No matter what amino acid sequence I start with and how many iterations I run, the 100 output structures from each iteration always end up with the SAME UNIQUE sequence. How to interpret this?

2) In the output sequence, there are many GLY residues, which doesn’t produce any meaning interaction with the enzyme. Are there any flags missing in my run?

3) The N- and C-termini are supposed to be covalently-bonded since it’s a CYCLIC peptide. However, the output from rosettaDesign always produce FREE termini. Is there any way to force the NT and CT to be connected?

Here are the flags I used:
-design -fixbb -read_all_chains -fa_input -repack -ex1aro -ex2aro_only -resfile myresfile -pdbout -ndrun 100

Thank you in advance,

Shirley

[Anonymous]

1) It means the packer really likes that sequence. How large is the sequence space you’re searching? How many rotamers? If it’s small you may be searching it exhaustively.

2) It probably means you’re getting a lot of sidechain clashes. Look up the soft_rep_design flag and see if it helps (it softens VDW repulsion).

[Anonymous]

> 1) It means the packer really likes that sequence. How large is the sequence space you’re searching? How many rotamers? If it’s small you may be searching it exhaustively.
>
> 2) It probably means you’re getting a lot of sidechain clashes. Look up the soft_rep_design flag and see if it helps (it softens VDW repulsion).
>
>

Thank smlewis very much for the comments. Here are the answers to your questions:

1) My cyclic peptide is 8-residue long. I would expect some sequence like F-E-K-A-R-V-R-R that satisfy surrounding environment (TRP, TYR, ASP, SER, GLN,…), but what I got is always P-G-G-G-D-G-G-G. How to find out “how many rotamers”?

2) The first sequence to start with is A-A-A-A-A-A-A-A, so I would think there shouldn’t be many sidechain clashes. I’m going to try “soft_rep_design”.

[bpierce]

You might also try minimizing the starting structure before running the design (then run design with regular weights and with soft_rep_design, separately).
A cmd line I like to use before doing any design on a crystal structure is:

rosetta.gcc -relax -farlx -minimize -sc_only -read_all_chains -use_pdb_numbering -multi_chain -use_input_sc -fa_input [-find_disulf -read_hetero_h2o] -ex1 -ex1aro -ex2 -s 1NAT.pdb [-decoystats] -fa_output -nstruct 1

(I put the optional flags in [ brackets ] ie not for every situation, though they won’t hurt anything).

[Anonymous]

> You might also try minimizing the starting structure before running the design (then run design with regular weights and with soft_rep_design, separately).
> A cmd line I like to use before doing any design on a crystal structure is:
>
> rosetta.gcc -relax -farlx -minimize -sc_only -read_all_chains -use_pdb_numbering -multi_chain -use_input_sc -fa_input [-find_disulf -read_hetero_h2o] -ex1 -ex1aro -ex2 -s 1NAT.pdb [-decoystats] -fa_output -nstruct 1
>
> (I put the optional flags in [ brackets ] ie not for every situation, though they won’t hurt anything).

Vanita, Thank you very much for sharing the trick. I’ll definitely try. I wonder whether it is possible to minimize part of a complex structure, say, the ligand only.

[Author]

Posts