Observable Reference
Reference for molecular observables computed by the NeuroCGMD analysis pipeline. Each observable is reported at the level where it is most interpretable — CG for collective properties and AA for coordinate-level structural analysis.
▶ STRUCTURAL OBSERVABLES (CG) — computed on CG bead positions
RMSD
Root Mean Square Deviation
Mass-weighted structural deviation from a reference configuration. Measures overall structural drift over the trajectory.
Units: nm. Computed at CG level. A plateau in RMSD indicates the system has reached structural equilibrium.
RMSD = √(∑ mi |ri − rref,i|² / ∑ mi)
Units: nm. Computed at CG level. A plateau in RMSD indicates the system has reached structural equilibrium.
RMSF
Root Mean Square Fluctuation
Per-particle time-averaged positional fluctuation. Identifies flexible loops, hinge regions, and rigid structural cores.
Units: nm. High RMSF beads correspond to flexible surface loops; low RMSF beads form the structural core.
RMSF(i) = √(<|ri − <ri>|²>)
Units: nm. High RMSF beads correspond to flexible surface loops; low RMSF beads form the structural core.
Radius of Gyration (Rg)
Mass-weighted compactness measure
Quantifies the overall size and compactness of the molecular system.
Units: nm. Decreasing Rg indicates compaction (e.g., binding). Stable Rg indicates structural integrity.
Rg = √(∑ mi |ri − rCOM|² / ∑ mi)
Units: nm. Decreasing Rg indicates compaction (e.g., binding). Stable Rg indicates structural integrity.
SASA
Solvent Accessible Surface Area
Shrake-Rupley algorithm with configurable probe radius (default 0.14 nm). Distributes test points uniformly on a sphere around each particle and counts exposed points.
Units: nm². Decreasing SASA upon binding indicates interface burial. Tracks solvent exposure changes during conformational transitions.
Units: nm². Decreasing SASA upon binding indicates interface burial. Tracks solvent exposure changes during conformational transitions.
thermodynamic observables
▶ THERMODYNAMIC OBSERVABLES (CG) — free energy, pair correlations, reaction coordinates
RDF g(r)
Radial Distribution Function
Pair correlation function describing the probability of finding a particle at distance r from another, normalized by ideal gas density.
Units: dimensionless. First peak position reveals equilibrium inter-particle distance. g(r) → 1 at large r indicates proper normalization. Computed from fixed simulation volume.
g(r) = <n(r)> / (4πr² Δr ρ)
Units: dimensionless. First peak position reveals equilibrium inter-particle distance. g(r) → 1 at large r indicates proper normalization. Computed from fixed simulation volume.
Potential of Mean Force
Free energy from Boltzmann inversion
Converts the probability distribution of a collective variable (typically COM distance) into a free energy profile:
Units: kT. Minima correspond to metastable states. Barriers indicate transition state energetics. Limited by sampling — regions not visited have undefined PMF.
W(r) = −kBT ln[P(r) / Pmax]
Units: kT. Minima correspond to metastable states. Barriers indicate transition state energetics. Limited by sampling — regions not visited have undefined PMF.
2D Free Energy Landscape
Two-dimensional PMF
Joint probability distribution of two collective variables (COM distance and Rg) converted to free energy:
Displayed as a heat map with trajectory overlay showing start and end points. Identifies binding pathways and metastable basins.
F(x,y) = −kBT ln[P(x,y) / Pmax]
Displayed as a heat map with trajectory overlay showing start and end points. Identifies binding pathways and metastable basins.
Reaction Coordinate
Center-of-mass distance
Mass-weighted center-of-mass distance between binding partners over time. Serves as the primary reaction coordinate for binding processes.
Units: nm. Decreasing distance indicates binding progression. Fluctuations reveal the dynamics of approach, contact, and separation events.
Units: nm. Decreasing distance indicates binding progression. Fluctuations reveal the dynamics of approach, contact, and separation events.
AA-level observables from back-mapped structures
▶ H-BOND & CONTACT ANALYSIS (AA) — requires actual amino acid identity and atomic geometry
H-bond Detection
Geometric criteria with angle validation
H-bonds are detected on back-mapped AA structures using combined distance and angle criteria to minimize false positives:
Distance: donor-acceptor < 3.5 Å
Angle: D-H···A > 120° (approximated from CA→donor direction)
Proline: excluded as donor (secondary amine, no H)
Backbone: N (donor) and O (acceptor)
Side-chain donors: ARG (NE, NH1, NH2), LYS (NZ), ASN (ND2), GLN (NE2), HIS (ND1, NE2), TRP (NE1), SER (OG), THR (OG1), TYR (OH), CYS (SG)
Side-chain acceptors: ASP (OD1, OD2), GLU (OE1, OE2), ASN (OD1), GLN (OE1), HIS (ND1, NE2), SER (OG), THR (OG1), TYR (OH)
Distance: donor-acceptor < 3.5 Å
Angle: D-H···A > 120° (approximated from CA→donor direction)
Proline: excluded as donor (secondary amine, no H)
Backbone: N (donor) and O (acceptor)
Side-chain donors: ARG (NE, NH1, NH2), LYS (NZ), ASN (ND2), GLN (NE2), HIS (ND1, NE2), TRP (NE1), SER (OG), THR (OG1), TYR (OH), CYS (SG)
Side-chain acceptors: ASP (OD1, OD2), GLU (OE1, OE2), ASN (OD1), GLN (OE1), HIS (ND1, NE2), SER (OG), THR (OG1), TYR (OH)
Contact Analysis
Heavy-atom closest approach
Residue-residue contacts detected from back-mapped AA heavy-atom distances:
Contact cutoff: 4.5 Å (heavy atoms, excludes H)
CA pre-filter: skip pairs with CA-CA > 12 Å (efficiency)
Persistent contact: present in > 50% of analyzed frames
Contacts are computed per amino-acid identity (for example, GLY31-HIS102) using residue names from the input structure, supporting clear interface inspection and comparative analysis.
Contact cutoff: 4.5 Å (heavy atoms, excludes H)
CA pre-filter: skip pairs with CA-CA > 12 Å (efficiency)
Persistent contact: present in > 50% of analyzed frames
Contacts are computed per amino-acid identity (for example, GLY31-HIS102) using residue names from the input structure, supporting clear interface inspection and comparative analysis.