Protein Dipole Moments Server

Clifford Felder and Joel Sussman, Dept. of Structural Biology
Weizmann Institute, 761000 Rehovot, Israel

With this server you can discover if your protein might have an unusually large net charge or dipole moment, and how this might relate to specific structural features of the protein, and thereby its function.

The results are displayed with jmol pictures, with the dipole and mass moment vectors as red and grey-green lines, respectively.

Warning: the quadrapole moment and mass moment vector (RG) are being removed from this server because their calculation is inaccurate or problematic.

For an overview of this server's purpose, please click here.

This server is described in Clifford E. Felder, Jaime Prilusky, Israel Silman, and Joel L. Sussman 2007, " A server and database for dipole moments of proteins", Nucleic Acids Research, 35, special Web Servers Issue.


Please enter PDB-ID Code of the protein you want to analize:

Alternatively, if the above box is blank, you can upload
or enter your own data in PDB format, at least the HEADER and ATOM records, and ending in TER and END RECORDS, here:

Click here to calculate the dipole for each peptide strand. . Click here to represent the dipole vector head by an arrow rather than a ball.

Optionally enter a string of up to 20 residues, separated by commas, for which you want the dipole angles calculated:

Click here to run calculation:


Click here to get full descriptions of the table captions and the the list of the average values and std. deviations derived from the PISCES 90% homology set of proteins from the OCA browser.

If you don't know the id-code for your protein, use the PDB Browser to find it.


Please note the following:

  1. The following information are returned: number of heavy atoms used, number of residues, mean radius of the protein, overall shape (spherical, spherical, prolate or oblate ellipsoidal or other), numbers of positively and negatively charges residues, net molecular charge, molecular dipole moment in Debyes, quadrapole moment, and the ratios of charge and of dipole moment over number of atoms.

  2. Besides providing this information for the particular protein requested, the server compares them against the corresponding average values from databases containing this information calculated for a set of over 10000 representative, non-homologous biological unit proteins and their constituent, unique (non-redundant) peptide chains, all with at least 50 residues, according to the PISCES list of nonhomologous proteins to the 90% homology level (See http://dunbrack.fccc.edu/PISCES.php), as obtained from the OCA browser at http://oca.weizmann.ac.il/

  3. What this server returns is the number of standard deviation units the given value for your protein is above (+) or below (-) the average value. While values close to 0. indicate an average behavior, those close to +/- 1. deviate significantly from the average. Values of +/- 2. or more deviate very significantly, and if it is the charge, dipole or quadrapole moment, indicates an unusual property that may have a special function.

  4. Only the non-hydrogen atoms of the standard amino acids of the protein itself are included in the dipole calculation. DNA, RNA, metals, hetero-atoms and groups and solvent are all ignored in this calculation.

  5. All of the peptide chains present explicitly in the specified PDB entry's ATOM records are used in the calculation. Symmetry-related chains or coordinates that comprise the complete X-ray crystallographic or biological structure are not included.

  6. The Parse set of partial atomic charges is used. To compensate the absence of backbone amino hydrogens HN in alpha helices, the charges of main chain atoms C and O are doubled.

  7. All GLU, ASP, LYS, ARG and C- and N-termini residues are 100% ionized, and all other residues are completely non-ionized. Note that N-terminus GLU and ASP and C-terminus LYS and ARG residues are considered uncharged, though they are Zwitterionic. N-terminus LYS and ARG and C-terminus GLU and ASP are counted like two residues.

  8. This server is limited to proteins up to 200,000 heavy protein atoms and 10000 residues; larger proteins will produce unpreditable results.

  9. An option is provided to analize the protein by individual peptide strands, instead of one overall calculation encompassing the entire complex of unique peptide strands.

  10. By definition the dipole vector points from the direction of net negative charge to the direction of net positive charge. In other words, you would put a (-) at the origin of the vector and a (+) at its end.

  11. You can also obtain the angle between the dipole moment vector and the Beta carbon of one or more residues that you can specify. This can let you know how the dipole moment lines up against key structural or binding regions of the protein, and might offer a clue to an electrostatic role assisting in the binding of certain substrates or inhibitors.

  12. The quadrapole moments are now calculated according to the formula, for each atom i, Qx(i) = 0.5q(i)*( 3x(i)**2 - r(i)**2 ), and similar for Qy and Qz, where q(i) is the atom's partial charge, x(i) is its X-coordinate and r(i) is its distance from the center-of-mass origin.

  13. In the pictures, the dipole vector is represented by a red line, and the truncated mass moment vector is a short line, half red and half black. Both vectors begin at the center-of-mass origin of the protein or chain. The origin of the red dipole line corresponds with the net negative charge, and the far end with the net positive charge of the dipole moment.

    Note that the non-peptidyl portions of the structure have been removed, and a new fictitious ligand representing the dipole and mass vectors has been added.

    In order for the jmol pictures to display, you must have java and javascript installed on your computer, and the java plugin installed in your web browser. For more information about jmol, see http://jmol.sourceforge.net/.

    Ordinary jmol pictures including the non-peptide moieties can be obtained from the RCSB PDB, OCA or FirstGlance servers.

  14. The drawn dipole vectors are scaled for large proteins. For short peptides, select the ligand and ask for smaller atom balls and thinner bonds.

  15. If you want to process a number of PDB entries using a script, or to link this server from another server and automatically have it come up with the results for a particular PDB entry, say entry 1abc, you can use URL
    http://bip.weizmann.ac.il/dipol-bin/dipol2j.cgi?pdb_id=1abc
    If you want to display by chains and calculate the angles for residues 10 and 20, use
    http://bip.weizmann.ac.il/dipol-bin/dipol2j.cgi?pdb_id=1abc&peps=y&iang=10,20

  16. If the PDB-ID code specified is not in the PDB because it is obsolete, the entry that replaces it will be used instead, automatically.


    If you have any questions, please feel free to email me at Clifford.Felder@weizmann.ac.il