Help Needed: Understanding mutational impact on Protein-DNA interactions via In Silico Docking with HADDOCK

HADDOCK
1

Hello everyone,

I am working on a project predicting deleterious mutations in a protein involved in gene expression regulation. I have performed mutations on chain A, conducted energy minimization, and subsequently performed protein-DNA docking using HADDOCK. Here are the results obtained (see attached table).

This is my first time using in silico methods, and I would appreciate assistance in analyzing and discussing these results. Which parameters should be considered significant for evaluating the impact of mutations? Do I need to interpret all parameters, particularly RMSD?

Thank you in advance for your help.

image
image1496×1174 56 KB

2

Dear AYMANE,

Thanks for your interest in using HADDOCK for you research.

  • First of all, if you already have a reference structure (probably the wide type one), you can make sure that results from your docking experiments are close (in term of RMSD) to that one.
  • Moreover, if you are using haddock3, you could simply use the alascan module, you could perform the mutation as well as obtain the difference in haddock score (and its components too) for each of your point mutations.
  • Also (possibly the most important one), please understand that for the several surveys and benchmarks we made, there is only a poor correlation between binding energy/delta G and the haddock score. Everything depends on the type of interactions you are looking at, and in some cases, the Electrostatic energy is more correlated to binding energy than other scores. Therefore, eventual conclusion obtained from the results of such experiments with HADDOCK should be considered with cautions.

To answer your question:

  • The RMSD value reported in the haddock2.4 webserver is the RMSD with respect to the best scoring model within one cluster, therefore evaluating the dispersion of positions with this set of complexes.
  • Using HADDOCK score is a good starting points, but evaluating the difference in Electrostatic energy could also be your best guess.
  • You must make sure that the complexes you obtain have the same conformation as the wild type, otherwise you cannot conclude anything about the scores.

With the hope that this answer will help you.

Cheers.

1 Like
3

Hello,

Thank you very much for your response. It has clarified several points for me.

Regarding your advice that I should ensure the results of my docking experiments are close to the reference structure in terms of RMSD, I have a further question. For example, in the table, the wild type has an RMSD of 0.5 +/- 0.3, while the R1401W variant has an RMSD of 9.1 +/- 0.0.

How can I ensure that the RMSD of this variant is close to that of the wild type? Should I simply repeat the docking of this variant until I obtain an RMSD of 0.5 +/- 0.3 or 0.4 +/- 0.3?

Thank you in advance for your help.

Best regards,

4

The RMSDs from the server are only telling you where a cluster is compared to the best model generated. They have no use/values in comparing different runs

1 Like
5

To ensure two structures are close, the best solution for now would be:

  • Use haddock3
  • Use a molecular viewer tool that can perform structural alignments and RMSD calculations (VMD, PyMOL, Chimera, VTX, etc…)
1 Like