# how to properly calculate the interaction between the QM and MM region

From: Oleksii Zdorevskyi (zdorevskyi_at_bitp.kiev.ua)
Date: Fri Oct 14 2022 - 04:59:58 CDT

Dear NAMD community,

I have a question regarding the calculation of interaction energies in
QM/MM simulations. I have browsed through previous NAMD-l discussions,
however, did not find any similar topics.

According to the NAMD QM/MM paper (Melo, Marcelo CR, et al. "NAMD goes
quantum: an integrative suite for hybrid simulations." Nature methods 15.5
(2018): 351-354.), the total *potential* energy of the system is computed
in the following way:

E(total) = E(MM) + E(QM) + E(QM/MM),

where E(MM) is the energy of the classical part, E(QM) - is the
single-point energy calculated by the QC software, and E(QM/MM) is the
interaction energy between the classical and quantum regions.

Now, I need to plot these 3 components as separate time series.

When I browse through the log file, I can see the "QMENERGY" keyword,
which corresponds to the E(QM) (actually, I checked it by comparing this
value with the corresponding single-point energy from ORCA output).

Then, I can extract the energy called "POTENTIAL" from the same log file.
If I understand correctly, this energy contains all the interactions in
the system, including van der Waals between the classical and quantum
part, as well as Coulomb interaction between classical charges, and
single-point charges derived from the previous QM step (when we use
"electrostatic embedding"). Does this energy describe E(tot), listed in
the formula above, correct?

If I subtract the "QMENERGY" (E(QM)) from the "POTENTIAL" (E(tot)), it
will give me the following:
E(MM)+E(QM/MM),

so E(MM) and E(QM/MM) can be only obtained as a sum.

However, I need to plot separate contributions for E(MM), E(QM), and
E(QM/MM).

Are there any means to extract them from NAMD output?

I will appreciate your help.

Best regards,
Oleksii

This archive was generated by hypermail 2.1.6 : Tue Dec 13 2022 - 14:32:44 CST