The Energy command calculates the potential energy of the selected atoms according to the current ForceField. This requires the soup to be clean (hydrogens added etc.) so that all force field parameters can be assigned.

Results are returned in units of either kJ/mol or kcal/mol, depending on the current EnergyUnit, and displayed both in the console and the simulation HUD on the right side. When doing interactive modeling, the energies can be updated quickly by pressing the energy key located in the top left corner, just left to the key '1' (on English keyboards it contains '~', on German keyboards '^').

Energy calculations are always done without a cutoff for electrostatic interactions . In a non-periodic box, the time required therefore grows with N^2, where N is the number of atoms. It is therefore usually faster to calculate energies in a periodic box, where the PME algorithm speeds things up, also if not explicitly selected with the Longrange command.

By choosing a command extension other than 'All', energies can be calculated individually for all units from atoms to objects (in all force fields but NOVA). The energy resulting from a pairwise interaction is distributed equally among the two interacting atoms. For angles and dihedrals, the two interacting atoms are the ones at the beginning and the end of the angle and dihedral, respectively. For improper dihedrals (planarity energy), the central atom receives the complete energy. Consequently, the sum of all individual energies equals the total energy, as shown in the following example:

# Load 1CRN
Clear
LoadPDB 1crn
Clean
ForceField Yamber3
Cell Auto
# Calculate complete energy
total = Energy
# Calculate per-atom energies
atomenergylist() = EnergyAtom all
print 'Energies (0.00+total) and (0.00+sum atomenergylist) are the same!'

Rules of thumb when working with force field energies:

• Energy calculations are more or less useless when done on a structure that has not been energy-minimized with the same force field before. If you take any PDB file, add hydrogens and calculate the energy, chances are good that you will obtain an arbitrary result. The reason is that your structure contains several bumps (interatomic distances much smaller than the sum of the Van der Waals radii), and bumps lead to artificially high force field energies due to the repulsive term in the Van der Waals energy component. Also bond lengths and angles are slightly off and add random components to the energy. The quickest way to solve this problem is to click on Options > Experiment > Energy minimization.

• The change of energy during a simulation is a matter of its own. When the simulation is started from an energy-minimized conformation (e.g. via the md_run macro), there is usually an increase in energy during the first picoseconds, since the added kinetic energy is partly stored as potential energy. On a larger time-scale, steady increases in energy can be observed if the simulation cell contains counter ions. These are initially placed at the positions with the lowest potential energy, usually close to charged solute groups, from where they detach to gain entropy, which in turn increases the potential energy (just like an NaCl crystal dissolves in water). Van der Waals energies get positive when many hydrogen bonds are present (e.g. in water), since they pull the atoms closer than their normal Van der Waals contact distance - which is however more than compensated by the gain in Coulomb energy.

• Contrary to quantum mechanics energies, force field energies can usually not be compared between different force fields. Only the general trend should be the same, e.g. a good protein structure assigned a strongly negative energy by force field A should also get a negative energy by force field B. Again, this requires that the protein has been energy minimized separately with both force fields before calculating the energy. When in doubt use the command 'Energy All', which prints all energy components individually, then you can quickly judge where unexpected differences come from. The most likely source of significant energy differences between force fields is the electrostatic Coulomb energy: while the NOVA force field includes electrostatic interactions between 1-4 bonded atoms in the Coulomb energy, the other force fields do not. This makes NOVA's Coulomb energy depend more strongly on the local interactions and thus the amino acid composition of a protein, e.g. proteins with many Ser/Thr residues tend to have a more positive Coulomb energy. This amino acid composition bias can be removed by subtracting the energy of the unfolded protein.

• When comparing force field energies, explicit solvent molecules add mainly random noise and should thus be replaced by an implicit solvent model. Also the comparison of point mutant energies after a molecular dynamics simulation is hardly possible by calculating energies for the entire protein, since side-chain movements far away from the actual mutation randomize the result. Use partial energy calculations to consider only the residues in the immediate neighborhood of the mutation, e.g. 'EnergyRes His 18' will calculate the energy contribution (and thus 'happyness') of a single residue.

• Force field energies help to judge the structural quality of a protein: distortions of local covalent geometry can be found by looking at the bond, angle and planarity energies. Unrealistically close contacts (bumps) lead to a high Van der Waals energy. On the other hand, the Coulomb energy is the least informative, because it depends significantly on the amino acid composition (e.g. proteins with a net charge have a higher Coulomb energy).

Energy

Calculate the complete force field energy of the soup.

Energy All

Calculate the individual energy components of the soup.

Energy VdW,Coulomb

Calculate the Van der Waals and Coulomb energy components of the soup.

bondnrg = Energy Bond

Assign the bond energy to variable 'bondnrg'.

EnergyRes Lys,Coulomb

Calculate the Coulomb energies of all Lys residues and display the results individually.

vdwlist() = EnergyAtom all,VdW

Calculate the per-atom Van der Waals energies and assign the results to the list 'vdwlist'.

bondanglist() = EnergyMol Protein,Bond,Angle

Calculate the bond and angle energies of all protein molecules and assign the results to the list 'bondanglist', which then contains twice as many elements as protein molecules present in the soup. First the bond energies, then the angle energies.