I want to calculate the band structure of a protein molecule. How can I do it using Siesta of Gaussian?

I am thinking about QE/VASP as the number of atoms are 600 and computationally it is not possible with QE/VASP. I am also not sure whether one can calculate if using any software or not.

Your suggestions will be highly appreciated.


1 Answer 1


Band structure is a concept for periodic system only. As proteins are not periodic structures, they don't have electronic bands.

The fact that you can download a PDB file (or a CIF file) with information about the crystal, it is just due to one of the ways we found to determine the protein structure that is crystalizing the protein and then using single crystal X-ray crystallography.

Very recently, a new technique called cryogenic electron microscopy (cryo-EM) is been using as another way to determine the protein structure without the need of passing through the crystallization process.

In case o non periodic system the molecular orbitals are used instead the electronic bands.

Even molecular orbitals are difficult to calculate for proteins due the high number of atoms (more than 1000 atoms), which make impracticable to use any DFT/ab initio software for the task.

But, you can use semi-empirical software like MOPAC that has implemented a function called MOZYME to do electronic calculations for proteins. Also, you can use the fragment molecular orbital method (FMO), implemented in software like GAMESS-US where you can use DFT/ab initio.

Another way is ti use Quantum Mechanic/Molecular Mechanic (QM/MM), where you divide your protein in two regions: one (smaller, few residues) where you use a quantum mechanic method and the other one (the rest of the protein) where you use molecular mechanics.

  • $\begingroup$ Thank you Sir for your details answer. I am looking for a similar study as reported in Fg. 3 "doi.org/10.1063/1.3584680" (band structure). But in this paper they have treated the protein as a periodic crystal. $\endgroup$
    – astha
    Commented May 2, 2021 at 12:02
  • 1
    $\begingroup$ @astha the paper you mention put the bases in a crystal, such that it becomes fully periodic. So it isn't a band-structure of a protein, rather a crystal of nucleotide bases. $\endgroup$
    – nickpapior
    Commented May 3, 2021 at 18:28
  • $\begingroup$ Thank you very much @Zeroth. Please if you do not mind, may I ask "How can I make my protein as a nucleotide bases"? $\endgroup$
    – astha
    Commented May 4, 2021 at 10:23
  • $\begingroup$ @astha, what you need is to look at sites like PDB to search for crystal structures with nucleutide bases. You cannot create a crystal of them by hand. $\endgroup$
    – Camps
    Commented May 4, 2021 at 14:15
  • $\begingroup$ Thank you Sir @Camps, you can find my files here as a tar file. It has three files. 1_* is taken from PDB, 2_* is cif file converted from 1_* using VNL (ATK-VNL) and 3_* is cif file created from 2_* using VESTA (just to keep it in proper format I opened 2_* in VESTA and exported as 3_*cif file. How can I know that the file 3_*cif is nucleotide bases of a protein file? please see the files here "we.tl/t-u8sEnHBAcj" . $\endgroup$
    – astha
    Commented May 11, 2021 at 5:40

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .