RNA-protin Docking starting with extended RNA structure

Hi All,

I am trying to use HADDOCK to find the structure of a complex composed of RNA and protein molecules based on restraints collected from NMR spectroscopy. I am using a set of unambiguous intermolecular restraints along with ambiguous restraints both derived from NMR and also based on AIR concept.

I have been quite successful getting the software to run locally starting with known structures of both molecules. However, the results are slightly deviating from the known x-ray structure of a homologous system. Also, I found that the results would be depending on the starting structure of individual molecules, especially the RNA molecule (x-ray structure of a homologous RNA vs. NMR derived structure, both are very similar to each other with RMSD of 2A).
That said, I wanted to try and make Haddock run starting with extended RNA structure and all NMR-derived restraints to guide the structure of RNA molecule. We have experimental restraints that we already used to calculate the RNA model, including unambiguous distances, H-bond basepair info as well as dihedral angles and sugar pucker.
I already tried to use all the aforementioned restraints to dock the complex starting with extended RNA, but that was quite unsuccessful. The calculated complex structure contains the exact same starting RNA structure, even though I kept the whole RNA segment fully flexible. Needless to say that all of my provided distance restraints are violated.

I am wondering if what I am trying to achieve is possible using Haddock2.2? and if I am missing something while setting up my run?

Your help greatly appreciated!


First of all, make sure to keep all hydrogen atoms if you have NOE-based distance restraints.

Second, I would not try to fold the RNA during the docking process, but first create confirmations that satisfy your NMR restraints and use those for docking.

You might indeed then define the molecules as fully flexibly, but make sure to include intramolecular restraints to maintain the 3D structure (for the protein you might also turn on the automatic dihedral angle restraints.

And possibly you will have to increase the number of steps for the it1 stage of HADDOCK

Dear Alexandre,

First of all thank you very much for your feedback and valuable instructions.
And I am sorry for replying you that late, I am new to Haddock and it took me a while to try couple of parameters.

It seems that starting with the folded RNA structure works much better indeed if I increase the number of steps in it1. I guess this is the only thing one need to keep in mind while trying to fold the protein/RNA while docking.

I still have a question though:
Is there a possibility to feed Haddock the base-pair planarity restraints? I have couple of them and wanted to make sure that those are included during folding. I was thinking that editing the “dan-rna_restraints.def” file to define them as Watson-Crick base-pairs and in the same time setting “basepair_planar=true” would do the job, but unfortunately all the it1 structures of this trial were failing.
Once more, I am using NMR based restraints that include protein-RNA restraints and RNA-RNA restraints from solid-state NMR, those include unambiguous distance restraints, h-bonds, and dihedral angles in addition to base-pair planarity information.

I am very sorry for the long message, I hope it is not getting very confusing.

Best regards,


Yes you can provide planarity restraints. For this to work you need indeed to edit the dna-rna_restraints.def file and then copy it into runX/data/sequence

Failures at it1 might come from various reasons. Look into some it1…out files and search for error messages starting at the bottom.

And if you define watson crick base pairs, these should be formed already otherwise it my give problems. The alternative would be to define additional distance restraints to force those.

I checked the .out file, it turned out that dan-rna_restraints file could not be read, my text editor added .txt to the file that is why Haddock could not locate it.

All my Watson-Crick pairs are already formed in the starting structure, and our restraints already contain those distances to keep them paired.

I will give it another shot and let you know the result.

By the way, what is the extent one need to increase the steps in it1 to achieve my task? I learned from another topic here that it need to increase to be at least four times of the standard ones, I am setting these to the following:
{===>} initiosteps=2000; rather than the standard 500
{===>} cool1_steps=2000; rather than the standard 500
{===>} cool2_steps=4000; rather than the standard 1000
{===>} cool3_steps=4000; rather than the standard 1000

Do you think this is enough, or do I need to change it to other values?

Best regards,

For your it1 question: You should try and see…