niccogreek/nmr-canonical-cleaned

Canonical NMR Dataset Collection — Data Card

Dataset release: v3
Canonical schema: v2
Spectral modalities: 1H and 13C resonance-level peak lists

Collection overview

This release brings several of the largest openly available processed NMR corpora used by current deep-learning methods into one model-independent schema. It combines simulated and literature-derived spectra while preserving the provenance and annotation coverage of every source.

The collection has five functional components:

1. train_val, the common representation-learning pool built from the MST-NMR, NMRexp, and NMRTrans/NMRSpec train and validation partitions;
2. test_benchmark, the union of their published test partitions after moving SimNMR-overlapping rich spectra into training and removing residual exact-canonical-SMILES overlap with the extended train pool and NMRGym;
3. ADMET subsets, exact molecule matches to three TDC endpoints, with the original TDC train/validation and test assignments;
4. SimNMR-PubChem / NMR-Solver, a much larger simulated shift-only component kept separate because its proton peaks contain shifts and equivalence-derived integration, while multiplicity, J coupling, and reported peak ranges are unavailable;
5. NMRGym, a smaller experimental shift-only component with paired 1H and 13C resonance lists and no supplied integration, multiplicity, J coupling, or reported peak ranges.

One Parquet row represents one spectrum record. A molecule may have multiple records when spectra originate from different sources, simulations, reports, or experimental conditions. Such records remain distinct unless their canonical structure and both exact shift lists are identical under the deduplication rule described below.

Source provenance

Component	Origin used by this collection	Reference
MST-NMR	Simulated paired spectra from the Multimodal Spectroscopic Dataset, using the processed NMRPeak LMDB release.	MSD, NMRPeak
NMRexp	Experimental records mined from chemistry Supporting Information published between 2010 and 2024, using the quality-controlled NMRPeak subset.	NMRexp, NMRPeak
NMRTrans / NMRSpec	Experimental peak tables mined from chemistry Supporting Information published between 2013 and 2025; this collection uses the released 212,440-record model dataset.	NMRTrans
SimNMR-PubChem / NMR-Solver	PubChem-scale simulated atom-level 1H and 13C shifts grouped through supplied equivalence classes.	NMR-Solver
NMRGym	Experimental paired 1H and 13C shift lists released as the scaffold-split NMRGym benchmark used by UltraNMR.	UltraNMR, source repository
ADMET labels	Ames, LD50 Zhu, and AqSolDB solubility endpoints with the official Therapeutics Data Commons splits.	Property benchmark notes

The source papers describe collection and upstream curation. This repository starts from the released processed representations and records the additional canonicalisation, split protection, and cleaning applied here.

Final files

Component	NMR records	Labels or derived features
Representation-learning pool	train_val.parquet	train_val_mol_properties.csv
Molecule-disjoint benchmark	test_benchmark.parquet	test_benchmark_mol_properties.csv
SimNMR-PubChem shift-only pool	simnmr.parquet	simnmr_mol_properties.csv
NMRGym experimental shift-only pool	nmrgym.parquet	nmrgym_mol_properties.csv
Ames	train_val.parquet, test.parquet	Matched train/validation and test labels
LD50 Zhu	train_val.parquet, test.parquet	Matched train/validation and test labels
AqSolDB solubility	train_val.parquet, test.parquet	Matched train/validation and test labels

The adjacent *_report.json files are compact processing reports generated by the same DVC run; admet/preparation_report.json records cohort construction.

Each molecular-property CSV has one row per final NMR record_id. It contains RDKit descriptors, functional-group indicators, ECFP4 and MACCS fingerprints, and an explicit RDKit processing status. Each ADMET CSV has the same unique record_id set as its paired Parquet and contains the TDC molecule, label Y, and source identifier.

Molecular-property CSV schema

The *_mol_properties.csv files are derived after the final filtering and disjoin steps. Their rows preserve the order of the paired Parquet and are joined through record_id; the values are calculated from smiles_canonical, not from the NMR peaks or a source property table. They are provided for analysis and structure-based baselines and are not acceptance criteria for the cleaned dataset.

Column	Meaning
record_id, smiles_canonical	Spectrum identifier and exact structure string used by RDKit.
rdkit_status	ok, missing_smiles, invalid_smiles, or calculation_error.
rdkit_error	Empty for ok; otherwise the retained processing failure.
exact_molecular_weight	RDKit exact isotopic molecular weight, in daltons.
calculated_logp	RDKit Crippen MolLogP estimate; dimensionless.
tpsa	RDKit topological polar surface area, in Ų.
hba, hbd	RDKit hydrogen-bond acceptor and donor counts.
rotatable_bonds	RDKit rotatable-bond count.
fraction_csp3	Fraction of carbon atoms that are sp3 hybridized.
aromatic_atom_fraction	Aromatic heavy atoms divided by all heavy atoms.
has_amine	1 when the SMARTS finds a trivalent amine N excluding amide/sulfonamide-like and imine N; otherwise 0.
has_amide	1 when an N–C(=O) amide pattern is present.
has_alcohol_or_phenol	1 when an alcohol or phenol O–H pattern is present.
has_ester	1 when a C(=O)–O–C ester pattern is present.
has_carboxylic_acid	1 when a C(=O)–OH carboxylic-acid pattern is present.
has_aldehyde_or_ketone	1 when an aldehyde or carbon-substituted ketone carbonyl pattern is present.
has_nitrile	1 when a C≡N pattern is present.
has_halogenated_group	1 when a carbon–F/Cl/Br/I bond is present.
has_heteroaromatic_ring	1 when an aromatic N, O, S, or P atom is present.
morgan_ecfp4_2048_hex	Radius-2, 2,048-bit Morgan/ECFP4 vector stored as 512 hexadecimal characters representing RDKit binary bytes.
maccs_keys_166_bits	The 166 usable RDKit MACCS positions as a 0/1 string; RDKit's unused position 0 is omitted.

The nine has_* fields are binary, potentially overlapping SMARTS indicators, not functional-group counts. For a non-ok row, all descriptor, indicator, and fingerprint cells are blank; the row itself is retained so Parquet/CSV alignment is not lost. The exact SMARTS definitions and fingerprint encodings are implemented in calculate_mol_properties.py.

Canonical schema

The physical representation is Arrow/Parquet with nested lists and structs:

CanonicalRecord
├── record and molecular provenance
├── atoms: list<string>
├── h_nmr_peaks: list<ProtonPeak>
│   ├── chemical shift and integration
│   ├── raw and harmonised multiplicity
│   ├── J-coupling list
│   ├── reported or reconstructed shift interval
│   └── optional equivalence-group provenance
└── c_nmr_peaks: list<CarbonPeak>
    └── shift plus optional source-specific simulated peak properties

The corresponding model-independent Python dataclasses are CanonicalRecord, ProtonPeak, and CarbonPeak. CanonicalParquetDataset streams one or more Parquet files as validated CanonicalRecord objects, while CanonicalNMRDataset provides the in-memory equivalent for small collections.

Record fields

Field	Arrow type	Meaning
record_id	string, required	Stable source-qualified spectrum identifier; unique within each file.
source	string	Canonical source family.
smiles	string	Structure string selected from the source release.
smiles_canonical	string	Isomeric canonical SMILES recalculated with RDKit.
molecular_formula	string	Formula recalculated from the canonical RDKit molecule.
nmr_frequency	string | null	Reported acquisition frequency when supplied.
nmr_solvent	string | null	Reported solvent when supplied.
atoms	list<string>	Atom symbols in RDKit canonical-SMILES order, normally heavy atoms.
h_nmr_peaks	list<struct<ProtonPeak>>	Canonical proton resonances; always a list.
c_nmr_peaks	list<struct<CarbonPeak>>	Canonical carbon resonances; always a list.

RDKit regenerates smiles_canonical, formula, and atom order from the selected source SMILES. This gives every source the same structural representation and records the RDKit version in the Parquet footer.

Proton peak fields

Field	Arrow type	Unit and convention
shift	float64, required	Chemical shift in ppm.
integration	int64 | null	Number of represented protons when supplied or derivable.
multiplicity_raw	string | null	Original source label.
multiplicity	string | null	Harmonised label used across rich sources.
j_values	list<float64> | null	J couplings in Hz.
range_min	float64 | null	Lower endpoint of the reported shift interval, in ppm.
range_max	float64 | null	Upper endpoint of the reported shift interval, in ppm.
range_half_span	float64 | null	(range_max - range_min) / 2, in ppm.
equivalence_class	int64 | null	Source equivalence-group identifier for atom-level simulated data.
member_shifts	list<float64> | null	Original atom-level shifts represented by one grouped resonance.

The canonical multiplicity vocabulary is: m, d, s, dd, t, ddd, q, dt, td, br, ddt, dq, tt, quint, dddd, qd, sept, ddp, ddq, bd, dqd. Lossless aliases are normalised as p → quint, hept → sept, and brd → bd. Other supplied labels become <unk>, while their exact source text remains available in multiplicity_raw.

Carbon peak fields

Field	Arrow type	Unit and convention
shift	float64, required	Chemical shift in ppm.
integral	float64 | null	Source-supplied simulated carbon integral.
intensity	float64 | null	Source-supplied simulated intensity.
width	float64 | null	Source-supplied simulated carbon width in ppm.

The three optional carbon quantities are populated by MST-NMR. Carbon peaks from NMRexp, NMRTrans, NMR-Solver, and NMRGym carry the common shift field.

Missing values

• The two modality columns are always lists. [] means that the record has no usable peaks for that nucleus.
• In the rich MST-NMR, NMRexp, and NMRTrans files, j_values is always a list. [] means that no numerical coupling is listed for that peak.
• Shift-only sources use null for annotations outside their source representation, including multiplicity, J values, and reported ranges.
• A numerical J=0 remains a supplied value and can be masked during model preparation.
• After common cleaning, all 16,733,130 proton peaks in the rich files retain populated positive integration, canonical multiplicity, and all three range fields.

Example streaming access:

from data.dataset import CanonicalParquetDataset

records = CanonicalParquetDataset("datasets/cleaned/train_val.parquet")

for record in records:
    h_shifts = [peak.shift for peak in record.h_nmr_peaks]
    c_shifts = [peak.shift for peak in record.c_nmr_peaks]

Run repository code with PYTHONPATH=scripts so the data package is available.

Source-specific representation

MST-NMR through NMRPeak

MST-NMR is the simulated rich source. Its NMRPeak release provides paired proton and carbon resonances. Proton centroid, nH, category, parsed J values, rangeMin, and rangeMax map directly to the shared fields. If centroid is absent, conversion uses delta, followed by the midpoint of the two range endpoints. A source peak containing only one shift receives a point interval:

range_min = range_max = shift
range_half_span = 0

MST-NMR also supplies the optional carbon integral, intensity, and width (ppm). This carbon width is a source-simulated peak property and is separate from the proton range_half_span.

Converter: convert_mst_nmr.py.

NMRexp through NMRPeak

NMRexp contains experimental spectra mined from literature and processed by NMRPeak. Its proton mapping is the same as MST-NMR, including point intervals for single reported shifts. Carbon records supply shifts only. Frequency and solvent are retained where they occur in the processed release.

NMRexp is the source with partial modality coverage: an experimental report may contain only 1H or only 13C. Those records remain useful and use an empty list for the unavailable modality.

Converter: convert_nmrexp.py.

NMRTrans / NMRSpec

NMRTrans uses a processed NMRSpec collection mined from chemistry-paper Supporting Information published between 2013 and 2025. The released proton token has the form:

[shift, range_half_span, multiplicity, integration, J_values]

The field called peak_width by the upstream loader is the half-span of the reported interval:

range_min = shift - range_half_span
range_max = shift + range_half_span
range_half_span = (range_max - range_min) / 2

Carbon values are shift lists. The release contains paired modalities and does not supply solvent or frequency fields.

Converter: convert_nmrtrans.py.

SimNMR-PubChem / NMR-Solver

SimNMR-PubChem is a simulated PubChem-scale shift database used by NMR-Solver. The source stores atom-level predictions in nmr_predict, element identities in atom_index, and supplied equivalence groups in equi_class. Conversion creates one resonance per nucleus and equivalence class:

shift = mean(member_shifts)
1H integration = number of hydrogen members

For proton resonances, equivalence_class and member_shifts preserve the atom-level source information. Multiplicity, J coupling, and reported ranges are represented as unavailable annotations. Carbon equivalence groups are also reduced to their mean shift. This shift-only representation is distributed as a separate component so models can use it for large-scale shift pretraining or a staged simulated-to-rich curriculum.

Converter: convert_nmrsolver.py.

NMRGym

NMRGym is an experimental shift-only benchmark released as train, validation, and test pickle files. Conversion preserves its paired 1H and 13C shift lists and concatenates the published partitions in their original order. Unlike SimNMR-PubChem, it does not supply proton equivalence groups or integration. Integration, multiplicity, J coupling, reported ranges, solvent, and frequency therefore remain null rather than being inferred. An absent modality would remain []; after common filtering every released NMRGym row has both modalities.

This component is distributed separately from the rich training pool so it can support experimental shift-only pretraining, domain adaptation, or explicit source-aware sampling.

Converter: convert_nmrgym.py.

Processing and split rationale

The release is produced by the explicit DVC graph in the project repository:

public raw releases
    → source-specific schema-v2 canonicalisation
    → rich train/validation and source-test merges
    → move SimNMR-overlapping rich test spectra into train/validation
    → remove residual benchmark overlap with extended train and NMRGym
    → common filtering and exact-shift deduplication
    → ADMET matching and removal from rich pretraining
    → aligned molecular properties and final analytics

All NMR-to-NMR comparisons use exact smiles_canonical equality. This keeps stereochemical distinctions represented by the canonical isomeric SMILES. ADMET structures instead use full RDKit InChIKeys because they originate from an external property collection.

1. Canonicalisation and rich merges

The five converters preserve source records in the common model-independent schema, recalculate molecular metadata with RDKit, validate records, and retain source provenance. The MST-NMR, NMRexp, and NMRTrans train/validation splits produce 1,875,334 rich records; their source test splits produce 208,431.

2. SimNMR overlap transfer

The benchmark is compared with canonical SimNMR-PubChem. The 112,197 rich test records representing 111,551 molecules also present in SimNMR are removed from test and appended to rich train/validation. This protects evaluation from the large shift-only pretraining pool without discarding the richer spectra. The extended rich train/validation pool therefore contains 1,987,531 records, and 96,234 records remain in the benchmark candidate pool.

3. Residual benchmark disjoin

The residual benchmark is compared with the extended rich train/validation pool and canonical NMRGym. This removes 7,901 records: 7,838 newly matched through the extended train pool and 63 through NMRGym. The resulting 88,333 records enter common filtering. The comparison script accepts one benchmark and any number of comparison datasets, while only removing rows from the benchmark.

4. Common quality filtering and deduplication

The same filter is applied to rich train/validation, benchmark, SimNMR, and NMRGym. A record is retained when at least one modality contains peaks and:

• shifts and J values are finite and within the documented physical ranges;
• each modality has at most 60 peaks and each proton peak at most six J values;
• supplied proton integrations are positive;
• the canonical structure is a single connected fragment.

Deduplication removes only identical canonical structure plus exact sorted 1H and 13C shift lists; no rounding or tolerance is used. Different spectra of the same molecule remain separate. Filtering removes 34,330 records from the extended rich pool and 436 from the benchmark, leaving 1,953,201 and 87,897 records respectively before ADMET removal.

5. ADMET preparation

Ames, LD50 Zhu, and AqSolDB solubility retain their official TDC split assignments. Repeated property identities are consolidated; discordant labels are excluded, and the single Ames identity occurring across official splits is removed from both splits. Full RDKit InChIKeys match the external structures to rich spectra. The union contains 5,401 matched molecules and removes 7,513 NMR records from pretraining, leaving 1,945,688 final train_val records.

6. Molecular properties, reports, and analytics

Molecular-property CSVs are calculated only after the final split, filtering, and ADMET operations. Every sidecar has the same row count, order, and record_id sequence as its paired Parquet. Compact JSON processing reports record stage inputs, outputs, counts, and useful reason breakdowns. The full pipeline, parameters, dependencies, and output hashes are tracked by dvc.yaml, params.yaml, and dvc.lock in the project repository.

Dataset-specific composition

Dataset	Source	Records	Unique canonical SMILES	With 1H	With 13C	With both
Train/validation	MST-NMR	773,008	772,951	773,008	773,008	773,008
Train/validation	NMRexp	995,124	995,124	870,029	849,287	724,192
Train/validation	NMRTrans / NMRSpec	177,556	177,556	177,556	177,556	177,556
Train/validation	All	1,945,688	1,872,304	1,820,593	1,799,851	1,674,756
Benchmark test	MST-NMR	3,717	3,717	3,717	3,717	3,717
Benchmark test	NMRexp	74,209	74,209	64,866	63,446	54,103
Benchmark test	NMRTrans / NMRSpec	9,971	9,971	9,971	9,971	9,971
Benchmark test	All	87,897	87,631	78,554	77,134	67,791

NMRexp accounts for the single-modality rich records. MST-NMR and NMRTrans remain paired in both files.

ADMET property subsets

Endpoint	Split	Records	Unique property SMILES
Ames	Train/validation	2,527	1,748
Ames	Test	455	334
LD50 Zhu	Train/validation	2,748	1,964
LD50 Zhu	Test	574	427
AqSolDB solubility	Train/validation	3,777	2,671
AqSolDB solubility	Test	770	567

A molecule can have several distinct NMR records, so spectrum-record counts can exceed unique property structures. Target definitions and units follow TDC.

SimNMR-PubChem shift-only component

The canonical source contains 105,764,812 simulated records. Common filtering removes 255,196, leaving 105,509,616 records and approximately 97,936,772 unique canonical SMILES. Both modalities are non-empty in 105,476,128 records.

NMRGym experimental shift-only component

The canonical source contains 269,999 records. Common filtering removes 4,904, leaving 265,095 records and exactly 265,095 unique canonical SMILES. Every retained record has both modalities.

Analytics

Detailed descriptive statistics, source comparisons, figures, annotation coverage, NMRGym and SimNMR shift-only analyses, and links to every generated CSV are collected in the analytics report. The report links every generated table and figure and documents how to reproduce the analysis from the project repository.

Use considerations

The primary learning unit is a structured resonance-level peak list linked to a molecular structure. The rich collection supports models that consume chemical shifts together with integration, multiplicity, ranges, and J couplings. SimNMR-PubChem supports simulated shift-set pretraining, while NMRGym supplies experimental shift-only examples for domain support and simulated-to-experimental curricula. The ADMET files support frozen-encoder probes and supervised property-prediction studies with exact spectrum-to-label alignment.

Literature-mined records vary in solvent, field strength, and reporting practice. Source-aware evaluation is therefore informative alongside aggregate metrics. Carbon integral, intensity, and width are MST-specific simulated quantities. RDKit canonicalisation standardises the supplied structure representation; salts, protonation states, tautomers, and missing stereochemical information retain the distinctions present in the source SMILES and identity rules described above.

Further implementation rationale and source audits are documented in Datasets.md, Canonicalization_Implementation_Notes.md, Multiplicity analysis.md, Dataset_Filtering_and_Processing.md, Dataset Analysis.md, and Properties Dataset.md.