Hugging Face's logo

Datasets:
Orbital-Materials
/
MofasaDB

Size:
100K<n<1M
Tags:
chemistry
materials
diffusion
Synthetic
License:
Dataset card Files
xet
Community
MofasaDB / README.md
AndersChri's picture
AndersChri
Update README.md
b5cb624 verified
preview code
|
raw
history blame contribute delete
20.2 kB
metadata
license: cc-by-4.0
tags:
  - chemistry
  - materials
  - diffusion
  - synthetic
pretty_name: MofasaDB
size_categories:
  - 100K<n<1M

MofasaDB

The MofasaDB is a publicly available dataset containing 200.000+ de novo generated MOF (Metal-Organic Framework) structures from Mofasa trained on QMOF (up to 170 atoms), along with their geometry-relaxed counterparts. The database is released alongside the paper Mofasa: A Step Change in Metal-Organic Framework Generation. A user-friendly web interface for search and discovery can be accessed at https://mofux.ai/.

Table of Contents

  1. Database Overview
  2. Quick Start
  3. Property Reference
  4. Structural Properties
  5. MOFID Properties
  6. Zeo++ Geometric Properties
  7. ORB Properties
  8. MOFChecker Properties
  9. MOF Fragment Properties
  10. Linker Properties
  11. Validation Metrics

Database Overview

The database contains unconditionally generated MOF structures from Mofasa, along with their geometry-relaxed counterparts.

Files

File Description
samples.db Original generated MOF structures
relaxed.db Geometry-relaxed versions of the samples
sample_latents/ ORB latent embeddings for samples
relaxed_latents/ ORB latent embeddings for relaxed structures

Data Alignment

The databases are row-aligned: row i in samples.db corresponds to row i in relaxed.db.

Indexing:

  • ASE databases are 1-indexed: first row is db.get(1)
  • NumPy arrays are 0-indexed: first element is array[0]
  • Therefore: latent[i] corresponds to db.get(i + 1)

Quick Start

Load a Structure 

from ase.db import connect

db = connect("samples.db")
row_id = 1
row = db.get(row_id)                    # Get first structure (1-indexed)
atoms = row.toatoms()              # Convert to ASE Atoms object
print(atoms.get_chemical_formula())

Access Properties 

# Get energy per atom
energy = row.data['properties']['orb_properties']['orb_energy_per_atom']

# Get pore diameter
lcd = row.data['properties']['pyzeo_geometric_properties']['lcd']

# Get topology (top-level property)
topology = row.data['topology']

Load Orb Latent Embeddings 

import numpy as np

latents = np.load("sample_latents/orb_latent_4_graph.npy")
latent = latents[row_id - 1]  # Convert 1-indexed row to 0-indexed array

Compare Sample and Relaxed 

sample_db = connect("samples.db")
relaxed_db = connect("relaxed.db")

# Row i in both databases correspond to the same structure
row_id = 100
sample_atoms = sample_db.get(row_id).toatoms()
relaxed_atoms = relaxed_db.get(row_id).toatoms()

print(f"Sample formula:  {sample_atoms.get_chemical_formula()}")
print(f"Relaxed formula: {relaxed_atoms.get_chemical_formula()}")

Handle Missing Data

Not all properties are available for every structure. Common causes include:

  • MOFID failure: If MOFID cannot identify the MOF building blocks (nodes, linkers, topology), these properties are set to "UNKNOWN", "ERROR", or empty lists for missing SMILES strings.
  • Zeo++ non-porous: If Zeo++ determines a structure has insufficient porosity for probe access, geometric properties (lcd, pld, accessible volume/surface area) may be missing, zero, or None.
  • Component absence: Latent embeddings for bound_solvent and free_solvent are zero vectors when structures contain no solvent molecules.

Property Reference

Properties are stored in row.data with nested paths. Some examples:

PROPERTY_PATHS = {
    # ORB model properties
    'orb_energy_per_atom': 'properties.orb_properties.orb_energy_per_atom',
    'orb_max_force': 'properties.orb_properties.orb_max_force',
    
    # Zeo++ geometric properties
    'lcd': 'properties.pyzeo_geometric_properties.lcd',
    'pld': 'properties.pyzeo_geometric_properties.pld',
    'dif': 'properties.pyzeo_geometric_properties.dif',
    'av_volume_fraction': 'properties.pyzeo_geometric_properties.av_volume_fraction',
    'av_cm3_per_g': 'properties.pyzeo_geometric_properties.av_cm3_per_g',
    'nav_volume_fraction': 'properties.pyzeo_geometric_properties.nav_volume_fraction',
    'asa_m2_per_g': 'properties.pyzeo_geometric_properties.asa_m2_per_g',
    'number_of_channels': 'properties.pyzeo_geometric_properties.number_of_channels',
    'number_of_pockets': 'properties.pyzeo_geometric_properties.number_of_pockets',
    
    # Crystal symmetry
    'spacegroup_number': 'properties.crystal_symmetry.symprec_0.01/spacegroup_number',
    'pointgroup': 'properties.crystal_symmetry.symprec_0.01/pointgroup',
    
    # MOFID properties
    'mofid': 'mofid',
    'mofkey': 'mofkey',
    'topology': 'topology',
    'smiles_nodes': 'smiles_nodes',
    'smiles_linkers': 'smiles_linkers',
    'cat': 'cat',
    
    # MOFChecker
    'mofchecker': 'properties.mofchecker',
    'mofchecker_valid': 'properties.mofchecker.mofchecker_valid',
}

Structural Properties

Lattice Parameters

Key Type Description
lattice_a float Unit cell length along the a-axis (Å)
lattice_b float Unit cell length along the b-axis (Å)
lattice_c float Unit cell length along the c-axis (Å)
lattice_alpha float Angle between b and c axes (degrees)
lattice_beta float Angle between a and c axes (degrees)
lattice_gamma float Angle between a and b axes (degrees)

Chemical Composition

Key Type Description
reduced_formula str Empirical (reduced) chemical formula of the structure

MOFID Properties

MOFID is a standardized identifier for MOF structures that encodes topology, nodes, linkers, and catenation information.

Key Type Description
mofid str Full MOFID identifier string. Format: {nodes}.{linkers} MOFid-v1.{topology}.cat{n}.
mofkey str MOFKey identifier (a hash-based representation of the MOF structure). Format: {hash}.{topology}.MOFkey-v1.{short_code}.
smiles_nodes str Concatenated SMILES strings of all distinct metal nodes (.-separated).
smiles_linkers str Concatenated SMILES strings of all distinct organic linkers (.-separated).
topology str Three-letter RCSR topology code (e.g., "pcu", "dia", "fcu").
topology_v2 str Alternative topology assignment (may differ from primary if ambiguous)
cat int Catenation number (degree of interpenetration). 0 = non-catenated, n = n-fold catenated

Crystal Symmetry

Computed using pymatgen's SpacegroupAnalyzer.

Key Type Description
spacegroup str Crystal system from space group analysis at symprec=0.01 (e.g., "cubic", "triclinic")
spacegroup_v2 str Crystal system from space group analysis at symprec=0.1 (more tolerant symmetry detection)

Detailed Crystal Symmetry (nested under properties.crystal_symmetry)

Key Type Description
symprec_0.01/pointgroup str Point group symbol (Hermann-Mauguin notation)
symprec_0.01/spacegroup str Space group symbol (Hermann-Mauguin notation)
symprec_0.01/spacegroup_number int International Tables space group number (1-230)
symprec_0.01/spacegroup_crystal str Crystal system name
symprec_0.1/pointgroup str Point group symbol (at looser tolerance)
symprec_0.1/spacegroup str Space group symbol (at looser tolerance)
symprec_0.1/spacegroup_number int Space group number (at looser tolerance)
symprec_0.1/spacegroup_crystal str Crystal system name (at looser tolerance)

Zeo++ Geometric Properties

Computed using Zeo++ via the pyzeo wrapper. These properties characterize the pore geometry and accessibility using a spherical probe (default: N₂ probe radius of 1.86 Å).

Pore Descriptors

Key Type Unit Description
lcd float Å Largest Cavity Diameter – Diameter of the largest sphere that can fit in the pore without overlapping framework atoms
pld float Å Pore Limiting Diameter – Diameter of the largest sphere that can percolate through the framework (i.e., the narrowest point along the largest channel)
dif float Å Diameter of Included sphere along Free path – Diameter of the largest sphere that can diffuse along the accessible path
number_of_channels int — Number of distinct connected channel systems in the framework
number_of_pockets int — Number of isolated pores (inaccessible to the probe molecule)

Volume Properties

Key Type Unit Description
av_volume_fraction float — Fraction of unit cell volume that is accessible to the probe
av_cm3_per_g float cm³/g Accessible pore volume per gram of framework
nav_volume_fraction float — Fraction of unit cell volume that is non-accessible (pocket volume)
nav_cm3_per_g float cm³/g Non-accessible volume per gram of framework
channel_volume_fraction float — Fraction of total void volume that belongs to channels
pocket_volume_fraction float — Fraction of total void volume that belongs to pockets

Surface Area Properties

Key Type Unit Description
asa_m2_per_cm3 float m²/cm³ Accessible surface area per unit volume
asa_m2_per_g float m²/g Accessible Surface Area per gram (comparable to BET surface area)
nasa_m2_per_cm3 float m²/cm³ Non-accessible surface area per unit volume
nasa_m2_per_g float m²/g Non-accessible surface area per gram
channel_surface_area_fraction float — Fraction of total surface area belonging to channels
pocket_surface_area_fraction float — Fraction of total surface area belonging to pockets

ORB Properties

Properties computed using the ORB machine-learned interatomic potential.

Energy and Forces

Key Type Unit Description
orb_energy_per_atom float eV/atom Total predicted potential energy divided by number of atoms
orb_max_force float eV/Å Maximum force magnitude on any atom in the structure

ORB Latent Embeddings

ORB latent embeddings are stored as NumPy files in the sample_latents/ and relaxed_latents/ directories.

File naming: orb_latent_{layer}_{component}.npy

File Pattern Shape Description
orb_latent_{0-4}_graph (N, 256) Graph-level pooled latent
orb_latent_{0-4}_nodes_and_bridges (N, 256) Mean-pooled over metal nodes
orb_latent_{0-4}_linkers (N, 256) Mean-pooled over organic linkers
orb_latent_{0-4}_bound_solvent (N, 256) Mean-pooled over bound solvents
orb_latent_{0-4}_free_solvent (N, 256) Mean-pooled over free solvents
  • Layers 0-4 correspond to different depths in the ORB GNN (layer 4 = final layer)
  • Zero vectors indicate missing data (e.g., structures without solvents)

MOFChecker Properties

Computed using MOFChecker, a tool for validating MOF structures. All keys are prefixed with mofchecker_.

Validity Checks (Binary)

These descriptors are used to determine overall MOF validity. True indicates a problem (except where noted).

Key Type Description
mofchecker_valid bool Overall validity flag. True if structure passes all validity checks.
mofchecker_no_carbon bool True if structure contains no carbon atoms (invalid for organic-based MOFs)
mofchecker_no_hydrogen bool True if structure contains no hydrogen atoms
mofchecker_no_metal bool True if structure contains no metal atoms
mofchecker_has_atomic_overlaps bool True if any atoms are too close together
mofchecker_has_lone_molecule bool True if structure contains disconnected molecular fragments
mofchecker_has_overcoordinated_c bool True if any carbon has too many bonds
mofchecker_has_overcoordinated_n bool True if any nitrogen has too many bonds
mofchecker_has_overcoordinated_h bool True if any hydrogen has too many bonds
mofchecker_has_undercoordinated_c bool True if any carbon has too few bonds
mofchecker_has_undercoordinated_n bool True if any nitrogen has too few bonds
mofchecker_has_undercoordinated_rare_earth bool True if any rare earth metal is undercoordinated
mofchecker_has_undercoordinated_alkali_alkaline bool True if any alkali/alkaline earth metal is undercoordinated
mofchecker_has_suspicious_terminal_oxo bool True if structure has potentially incorrect terminal oxo groups on metals
mofchecker_has_geometrically_exposed_metal bool True if any metal has unusual coordination geometry
mofchecker_has_high_charges bool True if computed partial charges are unusually high

Informative Checks (Binary, not used for validity)

Key Type Description
mofchecker_has_oms bool True if structure has Open Metal Sites (coordinatively unsaturated metals)
mofchecker_has_3d_connected_graph bool True if the framework is 3D-connected (expected for MOFs)

Structure Hashes

Key Type Description
mofchecker_graph_hash str Hash of the full structure graph (atoms + bonds)
mofchecker_undecorated_graph_hash str Hash of graph with hydrogen atoms removed
mofchecker_decorated_scaffold_hash str Hash of framework scaffold with decorations
mofchecker_undecorated_scaffold_hash str Hash of bare framework scaffold
mofchecker_symmetry_hash str Hash encoding symmetry information

MOF Fragment Properties

Properties of the decomposed MOF components (nodes, linkers, solvents). Stored under properties.mof_fragments.

Component Types

MOF structures are decomposed into four component types:

  • nodes_and_bridges: Metal nodes and bridging groups
  • linkers: Organic linker molecules
  • bound_solvent: Solvent molecules coordinated to metal centers
  • free_solvent: Unbound solvent molecules in pores

Fragment Formulas

Key Type Description
{component}_formulas List[str] Chemical formulas of each fragment of this component type

Example: nodes_and_bridges_formulas = ["Zn4O", "Zn4O"] for a structure with two identical zinc nodes

Linker SMILES

Key Type Description
linkers_smiles List[str] Full SMILES strings for each linker fragment, including stereochemistry and charges where applicable
linkers_simple_smiles List[str] Simplified SMILES (scaffold only, no stereochemistry). More robust for parsing but less chemically accurate

Linker Properties

Molecular descriptors and fingerprints for organic linker molecules. Stored under properties.linker_properties.

Morgan Fingerprints

Morgan (circular) fingerprints are stored as NumPy files. For similarity search, use the standardized versions.

File Description
linkers_morgan_ecfp4.npy ECFP4 (radius=2), 2048-bit
linkers_morgan_ecfp6.npy ECFP6 (radius=3), 2048-bit
linkers_morgan_ecfp4_standardized.npy ECFP4 from standardized molecules
linkers_morgan_ecfp6_standardized.npy ECFP6 from standardized molecules

Scalar metadata:

Key Type Description
linkers_smiles_used List[str] Which SMILES string was successfully parsed for each linker (original, fixed, or simple)
linkers_smiles_standardized List[str] Chemically standardized SMILES (neutralized, canonical tautomer)
linkers_morgan_count_sum List[int] Sum of Morgan fingerprint bit counts (molecular complexity proxy)
linkers_morgan_count_sum_max List[int] Maximum count in Morgan fingerprint (indicates highly represented substructures)
linkers_morgan_count_sum_standardized List[int] Sum of counts for standardized fingerprints
linkers_morgan_count_sum_max_standardized List[int] Maximum count for standardized fingerprints

Molecular Descriptors

Computed on standardized molecules using RDKit.

Key Type Description
linkers_rotatable_bonds List[int] Number of rotatable bonds per linker (flexibility metric)
linkers_ring_count List[int] Number of rings per linker

Coordination Site Descriptors

Counts of metal-coordinating functional groups (computed on as-parsed molecules).

Key Type Description
linkers_coordination_site_count List[int] Total number of potential metal coordination sites per linker
linkers_coordination_site_breakdown List[Dict] Breakdown by coordination site type
linkers_carboxylate_count List[int] Number of carboxylate groups (-COO⁻/-COOH)
linkers_pyridine_count List[int] Number of aromatic nitrogen sites
linkers_imidazole_n_count List[int] Number of imidazole/triazole NH groups
linkers_primary_amine_count List[int] Number of primary amine groups (-NH₂)
linkers_secondary_amine_count List[int] Number of secondary amine groups (-NH-)
linkers_tertiary_amine_count List[int] Number of tertiary amine groups (-N<)
linkers_phosphonate_count List[int] Number of phosphonate groups
linkers_sulfonate_count List[int] Number of sulfonate groups
linkers_phenolic_oh_count List[int] Number of phenolic hydroxyl groups
linkers_alcoholic_oh_count List[int] Number of alcoholic hydroxyl groups
linkers_thiol_count List[int] Number of thiol groups (-SH)
linkers_nitrile_count List[int] Number of nitrile groups (-C≡N)

Validation Metrics

Binary metrics used to assess structure quality.

Key Type Description
no_atom_too_close bool True if all interatomic distances are physically reasonable
smact_valid bool True if composition passes SMACT electronegativity/charge balance checks
reconstruction_failed bool True if structure reconstruction from latent space failed

License

CC-BY-4.0

References