Multiscale Calculations with ORCA

Computational chemistry broadly employs two major approaches: quantum mechanical (QM) methods and classical molecular mechanics (MM) methods. QM methods explicitly treat electronic structures, making them ideal for investigating reaction mechanisms. In contrast, MM methods approximate molecules using classical physics, where atoms are modeled as point masses with charges and bonded interactions are represented as springs. Although MM methods lack electronic resolution and cannot accurately simulate bond breaking or formation, they are highly effective for studying overall molecular geometry, conformational dynamics, and large-scale motions.

To leverage the strengths of both approaches, hybrid multiscale methods have been developed. These methods divide a chemical system into regions: the chemically active site is modeled with QM, while the surrounding environment is treated with MM. This combined approach offers a more accurate and computationally efficient way to simulate complex systems, outperforming pure QM or small-cluster models in capturing the system’s full behavior.

This guide presents a customized protocol for performing QM/MM and related multiscale calculations using the ORCA software.

ORCA offers three major multiscale modeling frameworks:

· QM/MM: The active site is treated using QM, while the rest of the system is modeled with MM.

· QM1/QM2: Two QM layers are used, with QM1 describing the reactive center at high accuracy and QM2 providing a lower-cost description of the nearby environment.

· QM1/QM2/MM: A three-layer scheme combining QM1, QM2, and MM, enabling detailed yet efficient modeling of large systems.

The MM region can be divided into two parts: the active MM region and the extended MM region (containing non-active atoms). The active MM region is free to relax during geometry optimization, whereas the extended region is included in the optimization but with constrained atomic positions.

For additional information on ORCA’s multiscale implementations, consult the ORCA manual and our publication:
📄 DOI: 10.1038/s41598-024-67468-x

To perform multiscale calculations in ORCA, the following files are required:

· An equilibrated PDB file

· A topology file

· An ORCA input file for multiscale setup

The subsequent sections describe how to prepare each of these components.

1. Prepare the Equilibrated PDB File

1.1 Generate a Starting PDB Structure

Begin with a clean PDB structure of your system:

· For enzymatic systems, follow our AMBER-Setup protocol.

· For organic reactions in non-aqueous solvents, see the PackMol Setup section.

1.2 Equilibrate the System

Equilibrate the structure using molecular dynamics (MD) as described in the AMBER-MD page or any other preferred MD software.

Important: If periodic boundary conditions (PBCs) were used during MD, you must re-image the coordinates to ensure all molecules are properly wrapped in the simulation box. Otherwise, solvent molecules may appear disordered or "evaporated" in visualizations. In AMBER, this is done via cpptraj. To re-image your final coordinate file (mdrest2) and generate a wrapped version:

cpptraj prmtop ptraj.in

Where ptraj.in contains:

trajin mdrest2

trajout mdrest2-wrap restart

image origin center familiar com :1

This recenters the system around the center of mass of residue 1 (you can change :1 to the appropriate residue).

Convert the wrapped restart file into a PDB format using:

ambpdb -p prmtop -c mdrest2-wrap > equilibrated_system.pdb

2. Prepare and Convert the Topology File for ORCA

2.1 Generate the Topology File

Use AMBER’s tleap module to generate the topology (prmtop) file, or use a similar module in your MD software if you employed a different program. If you've already run MD simulations, this file should already be available.

2.2 Convert to ORCA Format

Convert the topology file to ORCA’s force field format (.orcaff) using the orca_mm utility:

orca_mm -convff -AMBER prmtop

This generates the file prmtop.ORCAFF.prms, which is used in QM/MM and QM1/QM2/MM calculations.

Note: A topology file is not required for QM1/QM2 calculations.

3. Setup of Multiscale Calculations

To perform multiscale calculations in ORCA, you need to define:

· QM system: Atoms treated using DFT.

· QM2 system: Atoms treated using a semiempirical method (in QM1/QM2 and QM1/QM2/MM calculations).

· MM Active region: Atoms treated using an MM method (in QM/MM and QM1/QM2/MM calculations).

· Extended active region: Fixed atoms surrounding the MM active region (optionally in QM/MM and QM1/QM2/MM calculations).

Only atoms in these regions are considered in the multiscale calculations; the rest of the system is excluded.

Manually assigning hundreds of atoms can be tedious. To streamline this process, use our custom utility, pdbtoorca. This free tool (available for Windows, Linux, and macOS) automates the setup for ORCA multiscale calculations and can be downloaded from GitHub.

Please cite the corresponding paper if you use this program:📄 https://www.nature.com/articles/s41598-024-67468-x

3.1 Define the QM System (s1 File)

Create a text file (s1) listing atom indices from the equilibrated PDB that belong to the QM region.

· Use one line per index.

· For ranges, use hyphens (e.g., 1-31 includes atoms 1 to 31).

· The pdbtoorca program automatically converts these to zero-based indexing for ORCA.

Example for a small molecule system:

1-31

Example for an enzymatic system:

155-167

252-268

In this example, side chains of an ILE and TYR are included. ORCA automatically detects junctions and adds link atoms (e.g., HL atoms based on CA and CB positions).

3.2 Convert PDB to ORCA Format

Before proceeding, convert the equilibrated PDB file (equilibrated_system.pdb; generated in section ‎1.2) to ORCA-compatible format:

pdbtoorca <<EOF

equilibrated_system.pdb

orcapdb

equilibrated_system_orca.pdb

EOF

3.3 Generate ORCA Input Files with pdbtoorca

Use the following scripts to set up different multiscale schemes. The inputs define active region thicknesses, shell sizes, CPU usage, charge/multiplicity, etc.

🧪 QM/MM Setup

pdbtoorca <<EOF

equilibrated_system_orca.pdb

qmmm

6.0

3.0

qmmm.inp

yes

prmtop.ORCAFF.prms

EOF

🧪 QM1/QM2 Setup

pdbtoorca <<EOF

equilibrated_system_orca.pdb

qmqm2

3.0

qm1qm2.inp

yes

EOF

🧪 QM1/QM2/MM Setup

pdbtoorca <<EOF

equilibrated_system_orca.pdb

qmqm2mm

3.0

6.0

3.0

qmqm2mm.inp

yes

prmtop.ORCAFF.prms

EOF

3.4 (Optional) Modify the Generated ORCA Input File

You can manually adjust the generated .inp file as needed. For instance, to switch from electrostatic to mechanical embedding, add:

Embedding Mechanical

in the %qmmm block.

📝 Sample qmmm.inp File

! RIJCOSX TPSSh def2-SVP def2/J D3BJ TIGHTSCF Opt

!QMMM

%qmmm

ORCAFFFilename "../prmtop.ORCAFF7.3A.prms"

QMatoms {0:28 } end

Use_QM_InfoFromPDB false

Use_Active_InfoFromPDB false

# Extension shell not used in this example

end

%geom scan

B 13 15 = 1.60, 3.6, 10

end

%pal

nprocs 32

end

*pdbfile 0 1 orcapdb.pdb

4. Run the Multiscale Calculations

Finally, execute the ORCA command as shown below to perform the multiscale calculations. If you want to run the calculations on multiple CPU cores, specify the full path to the ORCA executable. The following three commands execute ORCA for the respective calculations—QM/MM, QM1/QM2, and QM1/QM2/MM—using their input files, which can be generated from the scripts above:

/opt/orca_6_0_1/orca qmmm.inp > qmmm.out

/opt/orca_6_0_1/orca qm1qm2.inp > qm1qm2.out

/opt/orca_6_0_1/orca qmqm2mm.inp > qmqm2mm.out

Note: Adjust the path if your ORCA installation is located elsewhere or if you are using multiple cores for parallel processing.