chg_fit selects the atomic-charge source independently of iqm and qm_compare. Correction defaults to none, which preserves the charges from the input mol2. NCAA defaults to resp; none is invalid because the PDB route has no reusable input charges. MetalAA remains a fixed RESP-only workflow and does not expose chg_fit.
chg_fit | Behavior |
|---|---|
none | Correction only. Preserve the input mol2 charges without creating a charge-fitting work directory. |
resp | Run Gaussian ESP calculations followed by two-stage RESP. chg_level defaults to HF/6-31G(d) for Correction and NCAA. |
| Charge-capable MLIP | Use AIMNet2 (aimnet2), AIMNet2-NSE (aimnet2nse), or MACE-POLAR (macepols, macepolm, or macepoll). A recognized MLIP must provide atomic charges; otherwise Parmfit reports a capability error. |
| Unrecognized value | Pass the value verbatim to antechamber -c, for example abcg2, bcc, or gas. Parmfit does not maintain an Antechamber charge-method whitelist. |
RESP charge jobs always use Gaussian and reuse qm_nproc and qm_mem; they run whether iqm is enabled or disabled. The charge-capable MLIP may differ from the main #model. If it matches the main model, Parmfit reuses the active calculator and its configured options. A different charge-capable MLIP is initialized once on the active device with its default options. Parmfit does not provide a separate options block for the charge MLIP.
TorsionFit refines proper torsion parameters in two stages. Stage 1 uses a spectral shared-group restrained linear least-squares solver: torsion instances that share the same chemical environment (atom types plus one-hop neighbors) are grouped and fitted together, with a spectral helper providing phase seeds. Stage 2 takes the Stage 1 parameters and performs direct global kPhi/phase optimization against the scan reference profile.
iqm controls whether external QM jobs provide reference-quality geometry and Hessian data. When iqm=false, all optimization, Hessian evaluation, and torsion reference work use the MAPLE MLIP calculator exclusively. When iqm=true, MAPLE runs QM reference jobs whose scope depends on the route.
iqm does not affect RESP
-Abinitio routes always run Gaussian ESP jobs for RESP charge fitting regardless of the iqm setting. iqm controls only the reference geometry, Hessian, and torsion-scan data, not the RESP charge pipeline.
QM RESP does not require iqm
+Setting iqm=false does not disable QM charge fitting. Correction and NCAA still run Gaussian ESP and RESP when chg_fit=resp, and MetalAA still runs its fixed RESP workflow. iqm controls only whether external QM supplies reference geometry, Hessian, and torsion data.
| Parameter | Correction | NCAA | MetalAA | |||
|---|---|---|---|---|---|---|
chg_level | — | RESP ESP (always active) | RESP ESP (always active) | |||
chg_level | HF/6-31G(d); RESP ESP when chg_fit=resp | HF/6-31G(d); RESP ESP when chg_fit=resp | PBE1PBE/def2SVP; RESP ESP in the fixed workflow | |||
opt_level | QM opt/opt_freq/constrained_opt when iqm=true | QM opt/opt_freq when iqm=true | QM opt_freq when iqm=true | |||
sp_level | QM single-point via qm_mode when iqm=true; defaults to opt_level | QM single-point via qm_mode when iqm=true; defaults to opt_level | Unused (no TorsionFit) | |||
| Parameter | Default | Description | ||||
mol2 | Required | Input topology and atom typing for method=correction. | ||||
cmo | 0 1 | Target molecular charge and multiplicity. If omitted, Correction uses charge/multiplicity already attached to the input structure, falling back to 0 1. | ||||
bonded | mseminario | Bond/angle update method: mseminario, seminario, or none. | ||||
vib_scale | 1.0 | Force constant scaling factor for Seminario/mSeminario Hessian-based fitting | ||||
chg_fit | none | Charge source: input mol2 charges, resp, a charge-capable MLIP, or a value passed to antechamber -c. | ||||
chg_level | HF/6-31G(d) | Gaussian ESP level used only when chg_fit=resp. | ||||
chg_route | Empty | Additional Gaussian route text for RESP ESP jobs. | ||||
iqm | false | Use external QM reference optimization/Hessian and torsion references when enabled. See the main Parmfit page for details. | ||||
torsion_bonds | Auto | Optional explicit center bonds. Otherwise MAPLE selects rotatable single bonds from the mol2 topology. |
| Residue | ASP | GLU | LYS | ARG | HIP | CYM |
|---|---|---|---|---|---|---|
| Formal charge | -1 | -1 | +1 | +1 | +1 | -1 |
The large-model charge is Q_large = cmo_charge + Σ formal_charge(protein residues). The optional third cmo value (oxy) controls the metal formal charge identity (e.g. Fe2+ vs Fe3+) but does not participate in the total charge.
Correct protein preprocessing is essential: a deprotonated cysteine must be named CYM (charge -1), not CYS (neutral), or the inferred large-model charge will be wrong.
+The charge in cmo refers to the metal ion plus any non-protein ligands or cofactors—it is not the total system charge and is not assigned to a coordinating amino acid. Each matched protein residue contributes the charge of its Amber graph template. The large-model charge is therefore Q_large = cmo_charge + Σ net_charge(matched protein residues).
The optional third cmo value (oxy) controls the metal formal-charge identity (e.g. Fe2+ vs Fe3+) but does not participate in the total charge. A backbone-only nonstandard amino acid has no complete Amber template and must be parameterized separately; its provisional zero must not be treated as an established metal-site charge.
MetalAA is RESP-only
+MetalAA does not expose chg_fit. It always fits large-model charges with Gaussian ESP and RESP. chg_level and chg_route control that calculation, while chgmod and fixchg_resids retain their standard-residue fixed-charge semantics.
| Parameter | Default | Description |
|---|---|---|
target | Required | PDB ion residue selector for the metal center. |
target | Required | Unique PDB selector for the central ion residue; see Target. |
cmo | 0 1 | <charge> <multiplicity> [oxidation]. Oxidation is optional and used for metal identity/formal charge handling. |
cfmol2 | None | One or more cofactor/ligand mol2 templates for nonprotein residues near the metal site. |
add_resid | None | Additional residues to force into the metal-site core. |
| Parameter | Default | Description |
|---|---|---|
target | Required | PDB residue selector for the residue to parameterize. |
target | Required | Unique PDB selector for the protein residue to parameterize; see Target. |
cmo | 0 1 | Total charge and multiplicity for the target residue model. |
rn | MOL | Output residue name used in prepin/frcmod and processed PDB files. |
prom | ff14SB | Protein force field model, commonly ff14SB or ff19SB. |
ionm | 12_6 | Ion parameter set name passed to tleap deployment; NCAA does not enforce a fixed whitelist. |
bonded | mseminario | Bond/angle refinement: mseminario, seminario, or none. |
vib_scale | 1.0 | Force constant scaling factor for Seminario/mSeminario Hessian-based fitting |
chg_level | HF/6-31G(d) | Gaussian level for RESP ESP generation. Always active regardless of iqm. |
iqm | false | Use external QM reference geometry/Hessian/torsion references beyond the RESP jobs. See the main Parmfit page for details. |
chg_fit | resp | Charge source: resp, a charge-capable MLIP, or a value passed to antechamber -c. none is invalid for NCAA. |
chg_level | HF/6-31G(d) | Gaussian ESP level used only when chg_fit=resp. |
chg_route | Empty | Additional Gaussian route text for RESP ESP jobs. |
iqm | false | Use external QM reference geometry/Hessian/torsion references. It does not select the charge route. See the main Parmfit page for details. |
Bond/angle and torsion refinement are controlled separately. Setting bonded=none keeps the initial BOND/ANGLE terms; setting torsionfit=false keeps the initial DIHE terms. When either refinement is enabled, atom types are automatically reassigned (MAPLE Z0/Z1, etc.).
NCAA fits the optimized α and β capped conformers. With chg_fit=resp, Gaussian ESP data from both conformers enter one joint two-stage RESP fit. A charge-capable MLIP or Antechamber method evaluates the two conformers separately, after which MAPLE applies their equal-weight per-atom average. For model and Antechamber routes, a final backbone adjustment closes the target-residue charge in the prepin to the cmo value.
chg_fit is independent of iqm. RESP always uses Gaussian with chg_level, chg_route, qm_nproc, and qm_mem. NCAA cannot use chg_fit=none because a PDB input does not supply a reusable set of atomic charges.
All charge-route files remain under <base>_work/ncaa/. The resulting charged capped mol2 and target.chg feed the prepin/GAFF2 template build; the same charges are carried into the Stage 0 parameter set and the TorsionFit baseline.
NCAA uses a center-bond filter that restricts torsion fitting to target-residue sidechain torsions only; backbone–backbone and cap torsion bonds are excluded. For residues with particularly complex sidechains, consider setting torsionfit=false and relying on the original GAFF2 torsion parameters, or use torsion_bonds to refine specific dihedrals selectively. NCAA disables torsion ensemble targets internally; see the main Parmfit page for scan and refinement parameter details.
When the NCAA residue is inserted as a custom template, MAPLE dynamically generates ProteinFF-GAFF2 cross-terms at both peptide boundaries: the head side (prev_C–target_N) and the tail side (target_C–next_N). These include exact BOND, ANGLE, and DIHE terms that connect the custom template atom types to the neighbouring standard protein residue atom types.
Checked-in Parmfit config: examples/parmfit/NCAA.parmfit. Use it from a MAPLE input with #parmfit(method=abinitio,input=NCAA.parmfit).
A simple NCAA job keeps commonly used route options in the regular MAPLE input. The target selects the noncanonical protein residue, and the PDB block supplies the complete structure.
+#model=uma(size=uma-s-1p1)
+#parmfit(method=abinitio,target=A123,rn=MOL,cmo=0 1,chg_fit=resp,bonded=mseminario,torsionfit=false)
+#device=gpu0
+
+PDB ncaa.pdbFor a longer setup, keep the MAPLE launcher short and move Parmfit parameters into an external .parmfit file. The launcher and external config must both use method=abinitio. The PDB remains in the MAPLE input; do not place pdb, input, or global MAPLE options in the external config.
#model=uma(size=uma-s-1p1)
+#parmfit(method=abinitio,input=NCAA.parmfit)
+#device=gpu0
+
+PDB ncaa.pdbExample external Parmfit config
<base>_work/<rn>_maple.prepin and <base>_work/<rn>_maple.frcmod: Refined MAPLE residue template and parameter files.<base>_work/<base>_ncaa_tleap.pdb and <base>_work/<base>_ncaa_tleap.in: Processed protein PDB and tleap input.<base>_work/<base>_ncaa.prmtop, <base>_work/<base>_ncaa.inpcrd, and solvated PDB outputs: tleap validation products.<base>_work/ncaa/: RESP, conformer, antechamber, prepgen, and torsion-refinement intermediates.<base>_work/ncaa/: Optimized conformers, charge-route files, charged capped mol2, target.chg, AmberTools template intermediates, and torsion-refinement data.chg_fit=none is unsupported because NCAA has no input mol2 charges to preserve.cmo charge.