Recent changes in reporting Haddock scores

HADDOCK
1

Hi Haddock team,

I’ve been using the HADDOCK web server for quite a while now on one protein-peptide system but have noticed recently a stark change in the reporting of the scores for my docked poses.

Normally I have been seeing highest scores around -120/-130 but recently the scores being reported in the summary for repeats of the same systems have been as low as -220.

After using the traditional scoring weights:

HS(itw) = E(vdw) + 0.2 E(elec) + E(desolv) + 0.1 E(AIR)

I have found that the HADDOCK scores are in line with what I have seen previously for these systems so was wondering if you have changed the weights of the scoring function being reported in the summary or changed the function overall?

Thanks,
Charlie

2

Dear Charlie,

We have not changed the scoring function nor the weights.
This changes can be due to:

  • Number of restraints
  • Number of charges on your peptide (influenced by electrostatics)

We are also reporting values using interactive graphs, which can tell you which component has the most effect.

You can also try to re-launch a previous run, to make sure you can reproduce (or not) the data you had before.

But it is true that we have updated our server to HADDOCK2.5 now, and maybe something happened, we can investigate.

Cheers

3

Scores are different for various complex types.

The size of the interface will also impact the score.
And of course how well you satisfy your restraints as well.
To get a pure “energetic” score you should subtract the restraint energy from the score

The scoring function has not changed. And haddock2.5 has exactly the same functionalities as 2.4. It is only the Python3 version of haddock2.4

4

Hi,

Thanks for the quick reply.

I did try and rerun an old job which was why I was suspicious of the change in score.

I have been using the interactive graphs to quickly see the breakdown of the score components but have been unable to reproduce the score being reported.

For example here is the best result from my most recent run:

REMARK FILENAME=“complex_190w.pdb0”
REMARK ===============================================================
REMARK HADDOCK run for complex
REMARK initial structure: complex_190.pdb
REMARK ===============================================================
REMARK total,bonds,angles,improper,dihe,vdw,elec,air,cdih,coup,rdcs,vean,dani,xpcs,rg
REMARK energies: -695.213, 0, 0, 0, 0, -62.8039, -177.625, 15.7979, 0, 0, 0, 0, 0, 0, 0
REMARK ===============================================================
REMARK bonds,angles,impropers,dihe,air,cdih,coup,rdcs,vean,dani,xpcs
REMARK rms-dev.: 0,0,0,0,9.501262E-02,0.497887,0, 0, 0, 0, 0
REMARK ===============================================================
REMARK air,cdih,coup,rdcs,vean,dani,xpcs
REMARK >0.3,>5,>1,>0,>5,>0.2,>0.2
REMARK violations.: 1, 0, 0, 0, 0, 0, 0
REMARK ===============================================================
REMARK CVpartition#,violations,rms
REMARK AIRs cross-validation: 1, 1, 0.387267
REMARK ===============================================================
REMARK NCS energy: 0
REMARK ===============================================================
REMARK Symmetry energy: 0
REMARK ===============================================================
REMARK Membrane restraining energy: 0
REMARK ===============================================================
REMARK Local cross-correlation: 0.0000
REMARK ===============================================================
REMARK Desolvation energy: -11.3704
REMARK Internal energy free molecules: -45791.1
REMARK Internal energy complex: -44486.3
REMARK Binding energy: 1053.06

The energies I am using to calculate the score are in bold and the resulting haddock score from manual calculation being: -62.8039 + (0.2x-177.625) - 11.3704 + (0.1x15.797) = -108.1196

However the reported score in the summary is -212.722 so I’m wondering if there is any reason for the large discrepancy?

5

Do you have a link to the server output?

6

Here is the link to the server summary page:

https://rascar.science.uu.nl/haddock2.4/result/6531069906/393692-adult_homology_model_bovine_g

7

Ah! You are running solvated docking!

In that case you also need to add the protein-water energies to your vdw and elec energies.
These only include the interfacial waters (those within 3.5A of both chains)

These are listed in the PDB file, e.g.:


REMARK water - chain_1: -276.308 -2.78075 -273.527
REMARK water - chain_2: -156.156 -15.2655 -140.89

The first number is the total, the second the vdw energy and the third the electrostatic energy.

8

That’s interesting I never actually ticked the solvated docking option the only parameters I changed from the default were:
turning on refinement with short MD in explicit solvent
turning on writing additional pdb’s with solvent molecules

I know solvated docking was off because sampling 180 degrees during EM is always on by default and is not allowed with solvated docking (I have never turned this option off).

Could it be that since the switch the server is noticing that there is solvent present so is reporting the score in this way instead?

9

Indeed - you turned on keepwater and this caused the interactions with water to be part of the scoring.

Kind of unintended parameter combination.