Dear,

Following are few points to keep in mind while selecting a protein structure from PDB

1. Resolution must be minimum possible ( This will ensure the better quality of protein structure.)

2. Domain completeness. Examine your PDB structure and confirm the under study domain full structure availability. Partial domain will lead to false interpretations.

3. Variant /Mutations. According to your case study look for whether your structure is a wild type or a mutant/variant. In case of a mutant structure requirement you may have to introduce required mutations manually and model them.

4. Side Chain Completeness ( is of secondary importance). Structures determined through old techniques might have (Not always) missing side chains due to flaw in tech or manual error. Right 3D confirmation of Side chains is critical in small ligand binding thus ensure their completeness. As a possible solution you may look for latest structure availability of the same.

5. Ligand /Crystalline Water / Co factor presence. To get out the right docking result removal (As per case study ) of these elements form PDB file is important.

Wishes.

thanks Rajnikant.

Thanks

:)

To the previous good answer i would also add;

Search and superimpose high resolution structures of your target and select a structure that is a good representative, i.e, not distorted too far away from the others by, say, a large ligand. The search, superposition and visualisation can be done for instance in Relibase (www.ccdc.cam.ac.uk), free for academics.

Sometimes it may be right to select more than one pdb model to work with e.g. if there are significant protein backbone shifts

Thanks sir

Also check for the duplicate atoms like these:

ATOM 1263 N AASP A 831 28.952 5.339 56.003 0.50 46.16 N

ATOM 1264 N BASP A 831 28.946 5.352 55.998 0.50 48.15 N

ATOM 1265 CA AASP A 831 29.643 5.954 54.868 0.50 41.99 C

ATOM 1266 CA BASP A 831 29.642 5.987 54.876 0.50 45.50 C

ATOM 1267 C AASP A 831 30.901 6.726 55.228 0.50 39.76 C

ATOM 1268 C BASP A 831 30.902 6.732 55.268 0.50 41.60 C

Higher the resolution more abundant they become.

What would be criteria if we get the model from i-tesser

We have looked at this problem extensively while generating the Iridium-HT dataset. Criteria to consider include global (R-factor, Rfree, DPI) and local (RSCC, RSR, occupancy) quality metrics, the existence of alternate conformations, whether residues from an adjacent protein molecule in the unit cell contributes to the crystallographic binding site (as then the crystallographic model is not biochemically relevant) as well as ligand quality metrics. More information can be found in the publication:

http://dx.doi.org/10.1016/j.drudis.2012.06.011

In addition to what Mr rajnikant provided above, I would strongly recommend to use model co-crystallized with inhibitor. such a model can be used to validate accuracy of prediction of the software being used in study. Accuracy is generally measured by ability of a software to reproduce the crystallographic confirmation of ligand. If resulting RMSDs are having lower values (typically values below zero) then system is said to be validated. Many journals asks for validation if docking is concerned in publication. Moreover, i would like to rephrase the statement made by Rajnikant regarding resolution. The first statement should be like "Value of resolution must be minimum possible". Since, lower values indicate high resolution and hence good quality models. Therefore, Model with resolution value of 1.2 Angstrom or below should always be preferred against models with higher resolution values like 2.0 or 2.5 etc. It would be a good idea to read the paper describing model before you proceed for docking, so that you will gain a good idea regarding functional aspects like mode of inhibition and other related issues.

Best luck...

Resolution is a measure of the QUANTITY of data gathered on a structure, not the quality of the model developed to fit that data. Quality criteria for how well the model coordinates fit that data are metrics like R-factor, Rfree and DPI and these are things you should pay attention to. Completeness of the data is an often overlooked but useful criterion. Accordingly Mr. Meshram is not correct when he says a higher resolution structure should always be preferred, the model structure that best fits the data should always be preferred.

Mr. Meshram also stated above the RMSD of a redocked ligand pose should be less than 0. This is a physical impossibility, the lowest deviation obtainable from any prediction is 0. A cursory look at the expression for RMSD will confirm this.

Yes... i do agree with Dr Paul Hawkins regarding RMSD. And its my mistake while stating RMSD value. It should be well below 2. But I still insist using high resolution models keeping an eye on R-Free and R-Factor values too. Thanks Dr . Paul for constructive comment.

Hi to all,

Thannks f having this discussion. I found it very helpful.

I am currently learning about docking. I would like to ask.

Regarding of what Dr Paul Hawkins said about the RMSD of redocked ligand pose; any suggested acceptable RMSD range for redocking purpose?

There is a commonly used criterion of 2 Angstrom RMSD as a maximum for redo king the cognate ligand. This is a reasonable cutoff if you bear in the limitations of RMSD: it is size dependent and it is not comparing your docked prediction with experimental data ( which is electron density).

I'm converting PDB from xml into MySQL, still working on it, but I'm writing some documentation during that that may help you in some points.

https://docs.google.com/document/d/1kmqhi17KvmeS7W4jQlMi68Ng4YmXRaAD4Q51DW06qjU/edit?usp=sharing

Rajnikant Namdeo ,

Should we remove the ligand molecule from the pdb for docking studies?

Besides that, let me know the procedure for pdb structure to use as the receptor in Discovery studio Control ligand.

I think that it tends to be common to remove the ligand from the pdb prior to docking, either because it could interfere in the docking or to offer a point of reference so you can compare the crystallized/initial pose to the docked pose. It varies from program to program, but I tend to remove the ligand just to be sure.

Do you guys have any paper(journals) or book for me to read? I needed some facts to support my statement in my writing...

thanks u all

my compound is a lengthy chain structure.if it is possible for docking.if I want to give docking study what type of protein  I  want to select...........

Sivakumar Periyasamy -dear ensure that your compound should not violate lipinski rule of 5. Molecular wt of ur test compound should be less than 500 daltons and for protein target first check your compund's pharmacological activity by PASS online activity predictor software.

You can evaluate the quality of a crystal available in PDB with this summary. This section provides a succinct "executive" summary of key quality indicators. If there should be serious issues with a structure, this would usually be evident from this summary. In this pag you can read about each criteria.

https://www.wwpdb.org/validation/2017/XrayValidationReportHelp#overall_quality

This is an article that you can read to understand alla of this.

Article Models of protein-ligand crystal structures: Trust, but verify

thanks for the professional information

In order to obtain accurate docking results you need to search for a protein structure with a chemically similar to your target compound, co-crystallized ligand. If this is not the case, the PDB structure with the highest enrichment value should be the most appropriate.

Depending on your Interest, check the mutational status of the PDB and lowest possible resolution