torchcell.datamodels

The pydantic schema of a torchcell experiment. A record is a typed genotype x environment -> phenotype experiment: Genotype holds a list of gene perturbations (for example SgaKanMxDeletionPerturbation), Environment holds the medium, an optional temperature and any environment perturbations, and each phenotype family (fitness, gene interaction, CalMorph morphology, microarray, RNA-seq and pseudobulk expression, metabolite, protein abundance, visual score, environment response and others) has its own phenotype, experiment and experiment-reference classes in schema. The classes derive from the strict bases in pydant (ModelStrict and ModelStrictArbitrary). Dataset loaders emit these objects, and the converter modules map one experiment type onto another (for example gene essentiality onto fitness). The package namespace re-exports a subset (torchcell.datamodels.__all__); every class below is listed under the module that defines it.

compound_identity

Shared, pure, offline compound-identity resolver (UI-2).

Classes

compound_identity.CompoundResolutionStatus	Typed outcome of a resolution attempt, ordered from usable to terminal.
compound_identity.CompoundIdentityRecord	One row of the pinned name->structure table (JSON round-trips this natively).
compound_identity.CompoundIdentityResolution	What a caller splats into Compound (only the identity-carrying fields).

Functions

compound_identity.normalize_compound_name	Lowercase + strip + conservative synonym canonicalization (documented).
compound_identity.resolve_compound_identity	Resolve a compound identity from the pinned table -- pure, offline.
compound_identity.inchikey_from_smiles	Derive a standard InChIKey from a SMILES with RDKit, or None if it will not parse.
compound_identity.resolve_compound_identity_from_smiles	Resolve by name, falling back to the caller's SMILES ONLY when no row exists.
compound_identity.resolved_compound	Build a Compound fill-or-gap through the resolver (the loader entrypoint).

compound_identity_curate

Reproducible curator for the pinned compound-identity table (UI-2, serve-50).

Classes

compound_identity_curate.CurationDirective	One parsed input line: a loader label plus how to resolve it.
compound_identity_curate.PubChemProperty	One row of a PUG REST PropertyTable.
compound_identity_curate.PubChemClient	Rate-limited PUG REST client with an on-disk response cache.
compound_identity_curate.CuratedRow	One assembled table row, before it is serialized to JSON.

Functions

compound_identity_curate.parse_line	Parse LABEL [| key=value]... into a typed directive.
compound_identity_curate.read_name_list	Parse an input list of labels (comments and blank lines skipped).
compound_identity_curate.read_cid_list	Parse an input list of bare PubChem CIDs (one per line).
compound_identity_curate.name_property_url	The canonical single-name GET endpoint (also the row's recorded source_url).
compound_identity_curate.cid_property_url	The canonical single-CID GET endpoint that reproduces one batched row.
compound_identity_curate.synonyms_url	The canonical single-CID synonyms endpoint (the ChEBI source).
compound_identity_curate.assemble	Collapse directives + PubChem answers into deduplicated, canonical rows.
compound_identity_curate.serialize	Deterministic table bytes (sorted keys, stable indent, trailing newline).
compound_identity_curate.curate	Read the input lists, query PubChem, and assemble the rows.

conversion

Convert raw Neo4j experiment records into typed experiments stored in LMDB.

Classes

conversion.ConversionEntry	One input/output type pair plus its conversion functions.
conversion.ConversionMap	Collection of conversion entries defining how to map experiment types.
conversion.Converter	Abstract base that converts queried records and caches them in LMDB.

fitness_composite_conversion

Converter that maps gene essentiality and synthetic lethality data into fitness.

Classes

fitness_composite_conversion.CompositeFitnessConverter

Converter combining gene essentiality and synthetic lethality to fitness.

gene_essentiality_to_fitness_conversion

Convert gene-essentiality experiments into equivalent fitness experiments.

Classes

gene_essentiality_to_fitness_conversion.GeneEssentialityToFitnessConverter

Converter turning gene-essentiality records into fitness records.

Functions

gene_essentiality_to_fitness_conversion.gene_essentiality_to_fitness_experiment	Map an essential gene to a zero-fitness experiment, else drop it.
gene_essentiality_to_fitness_conversion.gene_essentiality_to_fitness_reference	Build the wild-type (fitness 1.0) reference for the fitness dataset.

identity

Composition-based identity for the environment-side entities of the graph.

Functions

identity.compound_identity_key	The join key for a compound, as "<field>:<value>".
identity.media_identity	Project a medium onto its composition: state, synthetic flag, base, contents.
identity.temperature_identity	Project a temperature onto its value and typed unit (no unit conversion).
identity.environment_perturbation_identity	Project an environment perturbation onto its typed slots.
identity.environment_identity	Project an environment onto medium, temperature, edits, oxygen, duration and rate.
identity.strain_background_identity	Project a strain background onto its genome content, dropping who stated it.
identity.identity_sha256	Hash an identity projection: sha256 over its canonical JSON encoding.

interned_constant

Content-addressed sub-objects of an Experiment record in the knowledge graph.

Functions

interned_constant.constant_id	The node id of an interned constant: sha256 over label, kind and exact payload.
interned_constant.split_experiment_dump	Return (pointered dump, constants) for one experiment.model_dump().
interned_constant.collect_pointers	Add every $ref id reachable in obj to out.
interned_constant.resolve_pointers	Splice constants back: a dict carrying $ref becomes constants[ref].
interned_constant.verified_constant	Parse a fetched payload after checking that (kind, payload) hashes to its id.

media

Reusable, provenance-first definitions of the common growth media.

Functions

media.dropout	base with the named compounds moved from components into dropouts.
media.restated	base with a dataset paper's own medium statements appended to its provenance.

ontology_checks

Programmatic coherence checks over the torchcell ontology.

Classes

ontology_checks.AdapterNodeSite	One BioCypherNode(...) construction in cell_adapter.py.
ontology_checks.PropertyMismatch	A node class whose emitted properties differ from its declared properties.
ontology_checks.OptionalTraversal	An adapter attribute chain that walks THROUGH a field the schema may null.
ontology_checks.PhenotypeLabelMap	Bijection between concrete phenotype classes and phenotype node classes.
ontology_checks.CompoundIdentityIssue	A Compound that is neither identified nor honestly gapped.
ontology_checks.MediaBaseIssue	A Media.base_medium that names nothing in MEDIA_LIBRARY.
ontology_checks.MediaDerivationIssue	A derived medium that dropped a base component without declaring a dropout.
ontology_checks.EnvironmentCompoundCollision	One compound encoded twice in an environment: as a component AND as an edit.
ontology_checks.JoinKeyCensus	How many records of a built dataset are joinable on each identity key.

Functions

ontology_checks.schema_models	Every pydantic model DEFINED in schema.py (not merely imported).
ontology_checks.schema_enums	Every enum DEFINED in schema.py.
ontology_checks.composition_targets	Schema models this model COMPOSES (holds in one of its fields).
ontology_checks.experiment_closure	Models reachable from Experiment / ExperimentReference.
ontology_checks.orphan_models	Schema models no experiment record can ever reach, sorted by name.
ontology_checks.lane_membership	The three record lanes, each closed under inheritance AND composition.
ontology_checks.lane_back_edges	Composition edges pointing UP from a lane into the experiment lane.
ontology_checks.sourced_value_fields	Model.field -> rendered annotation for every field carrying SourcedValue.
ontology_checks.enum_value_collisions	Schema enums whose members share a value (a silent Python alias).
ontology_checks.graph_schema	The committed BioCypher graph schema, parsed into per-class entries.
ontology_checks.cell_adapter_source	The CellAdapter module source (read as text; never imported for this).
ontology_checks.isolated_graph_node_classes	Node classes no edge class connects (invisible to any traversal query).
ontology_checks.phenotype_label_map	Match each phenotype node class to its concrete pydantic phenotype class.
ontology_checks.adapter_node_sites	Every BioCypherNode(...) the adapter constructs, with what it emits.
ontology_checks.adapter_property_mismatches	Node classes where emitted properties != declared properties, either way.
ontology_checks.adapter_optional_traversals	Adapter chains that dereference THROUGH a schema field that may be None.
ontology_checks.compound_has_identity	True when the compound carries at least one structure identifier.
ontology_checks.compound_has_identity_gap	True when an identifier's absence is declared as a typed ProvenanceGap.
ontology_checks.media_library_compound_issues	Shared-library compounds that are neither identified nor honestly gapped.
ontology_checks.media_base_issues	Library media whose base_medium names no member of MEDIA_LIBRARY.
ontology_checks.media_derivation_issues	Derived media that silently lost a base component.
ontology_checks.environment_compound_collisions	Compounds encoded twice in one environment: as a component AND as an edit.
ontology_checks.join_key_audit	Census of the identity keys a built dataset actually carries.

pydant

Strict Pydantic base models used across torchcell data schemas.

Classes

pydant.ModelStrict	Base model that forbids extra fields and is immutable.
pydant.ModelStrictArbitrary	Strict immutable model that also permits arbitrary field types.

schema

Pydantic data models for torchcell genotypes, environments, and phenotypes.

Classes

schema.ProvenanceGapMixin	Mixin giving a model a provenance_gaps list of typed field-absences.
schema.HashableProvenanceGapMixin	ProvenanceGapMixin for a model that must stay hashable.
schema.MatingType	Mating-type locus state.
schema.Zygosity	How many of a background's chromosome copies carry an allele.
schema.AlleleEdit	How a background allele's sequence differs from the R64 locus.
schema.GenomicSpan	A 1-based, end-inclusive interval on one replicon of a deposited assembly set.
schema.BackgroundAllele	One allele a strain background carries, as an edit against its reference genome.
schema.StrainBackground	The genome content a strain carries beyond its reference genome.
schema.ReferenceGenome	The deposited reference a record's identifiers and background edits are written against.
schema.StrainReferenceGenome	A ReferenceGenome that states its typed strain background (#507).
schema.SOTerm	A Sequence Ontology term: a SO:NNNNNNN id paired with its name.
schema.GenePerturbation	Base perturbation of a single gene by systematic and common name.
schema.PresenceAbsencePerturbation	AXIS 1 -- a gene is PRESENT or ABSENT relative to the reference.
schema.SequencePerturbation	AXIS 3 -- a sequence-level allelic change (SNP/indel/substitution allele).
schema.ExpressionRangeMultiplier	Min/max multiplier bounds on a gene's expression level.
schema.CrisprConstruct	The engineered CRISPR machinery introduced to perturb a gene -- a first-class MATERIAL ENTITY (the guide RNA + Cas effector we actually put in the cell).
schema.StrainConstruction	Where one physical deletion strain came from in its collection.
schema.OrfHistoryRelation	How the ORF a strain was built against relates to the current R64 gene.
schema.ConstructedOrf	The ORF annotation a deletion strain was BUILT against, when not the current gene.
schema.DeletionPerturbation	Gene deletion via KanMX or NatMX gene replacement (engineered absence).
schema.KanMxDeletionPerturbation	Gene deletion via KanMX gene replacement.
schema.BarcodedKanMxDeletionPerturbation	A KanMX deletion strain that carries the molecular barcode it was screened by.
schema.HeterozygousDeletionPerturbation	One allele of a diploid replaced by a deletion cassette (HIP / het collection).
schema.NatMxDeletionPerturbation	Gene deletion via NatMX gene replacement.
schema.SgaKanMxDeletionPerturbation	KanMX deletion perturbation specific to SGA experiments.
schema.SgaNatMxDeletionPerturbation	NatMX deletion perturbation specific to SGA experiments.
schema.DampPerturbation	Decreased-abundance-by-mRNA-perturbation (DAmP) allele perturbation.
schema.SgaDampPerturbation	DAmP perturbation specific to SGA experiments.
schema.TsAllelePerturbation	Temperature-sensitive allele perturbation via amino acid substitution.
schema.AllelePerturbation	Generic allele perturbation via amino acid substitution.
schema.SuppressorAllelePerturbation	Suppressor allele that raises fitness in the presence of a perturbation.
schema.SgaSuppressorAllelePerturbation	Suppressor allele perturbation specific to SGA experiments.
schema.SgaTsAllelePerturbation	Temperature-sensitive allele perturbation specific to SGA experiments.
schema.SgaAllelePerturbation	Generic allele perturbation specific to SGA experiments.
schema.ConditionalAlleleClass	How a conditional allele of an essential gene reduces its function.
schema.ConditionalAllelePerturbation	A conditional (hypomorphic) allele of an essential gene, class possibly unknown.
schema.MeanDeletionPerturbation	Deletion perturbation aggregating duplicate experiments by their mean.
schema.MarkerDeletionPerturbation	Gene deletion via a selectable/auxotrophic marker other than KanMX or NatMX.
schema.CrisprDeletionPerturbation	Gene deletion via an ACTIVE Cas nuclease cut repaired from a homology donor.
schema.GeneAdditionPerturbation	Gain-of-function: a gene ADDED to the strain (heterologous expression or an extra native copy), carried on a plasmid or integrated at a chromosomal locus.
schema.NaturalGeneAbsencePerturbation	A reference (core) gene ABSENT in a natural isolate (Caudal core-loss).
schema.NaturalGenePresencePerturbation	A non-reference (accessory) gene PRESENT in a natural isolate (Caudal accessory).
schema.SequenceVariantPerturbation	SNP/indel-level allelic variation: a native gene whose sequence in this strain differs from the S288C reference, captured via an off-graph pointer.
schema.CopyNumberVariantPerturbation	Copy-number variation (CNV) of a PRESENT pangenome ORF relative to S288C.
schema.EngineeredCopyNumberPerturbation	An ENGINEERED copy-number/dosage change of a PRESENT native gene.
schema.ExpressionModulationPerturbation	AXIS 4 (ABC) -- engineered expression modulation of a present gene.
schema.CrisprActivationPerturbation	CRISPRa -- guide-directed dead-Cas activator INCREASES a present gene's expression.
schema.CrisprInterferencePerturbation	CRISPRi -- guide-directed dead-Cas repressor DECREASES a present gene's expression.
schema.Genotype	Collection of gene perturbations defining a strain's genotype.
schema.TemperatureUnit	UO-aligned temperature units (typed, not a free string) -- G2 unit typing.
schema.Temperature	Temperature value with a typed unit (defaults to Celsius).
schema.ConcentrationUnit	UO-aligned units for a dose / physical-factor magnitude (G2 unit typing).
schema.DoseBasis	How a dose was SET when the molar value is not released (dose PROVENANCE, not a phenotypic consequence): a target-inhibition endpoint or an explicit fixed dose.
schema.PhysicalFactor	A neutral, scalar physical/physiological environment variable.
schema.Compound	Chemical identity of a small molecule, keyed by a canonical InChIKey.
schema.Concentration	A dose of an environmental agent: a numeric value+unit and/or a target-basis.
schema.Solvent	The vehicle a compound was dissolved in and its final fraction in the medium.
schema.MediaComponentRole	Functional role of a medium component (never a phenotypic consequence).
schema.ComponentDefinition	Identity/composition completeness of a component (the 'what is it' axis).
schema.MediaComponent	One ingredient of a medium: a (possibly undefined) compound at a dose.
schema.Media	Growth medium resolved to typed COMPONENTS (provenance-first).
schema.EnvironmentPerturbation	Base: a defined change to the growth environment vs the base medium.
schema.SmallMoleculePerturbation	An added chemical SPECIES (drug, acid, alcohol, salt, oxidant, ...).
schema.EnvironmentPhysicalPerturbation	A neutral, scalar physical/physiological environment FACTOR (pH, osmolarity, ...).
schema.BiologicAgentClass	Material class of a proteinaceous / biologic agent added to the medium.
schema.BiologicPerturbation	An added BIOLOGIC agent (peptide / protein / antibody / toxin).
schema.EndpointRule	When a culture was read out (the endpoint is part of what was measured).
schema.CultureFormat	The physical culture a measurement was grown in.
schema.PreCultureSource	What the screened culture was inoculated FROM.
schema.PreCulture	The culture step BEFORE treatment, which sets the cells' state at time zero.
schema.Environment	Experimental environment: base medium + temperature + optional perturbations.
schema.CultureEnvironment	An Environment that also states its culture protocol (#507 point 8).
schema.Phenotype	Base phenotype describing an observed label and its graph level.
schema.UncertaintyType	What a reported uncertainty number IS, so it converts to an SE correctly.
schema.SampleUnit	What one sample in n_samples physically is.
schema.FitnessPhenotype	Fitness phenotype with fitness value and uncertainty statistics.
schema.GeneEssentialityPhenotype	Phenotype indicating whether a gene knockout is lethal.
schema.SyntheticLethalityPhenotype	Phenotype indicating synthetic lethality between perturbed genes.
schema.SyntheticRescuePhenotype	Phenotype indicating one perturbation rescues another's deleterious effect.
schema.GeneInteractionPhenotype	Phenotype holding a gene interaction score, its p-value and its replicate design.
schema.CalMorphPhenotype	Phenotype holding CalMorph morphological measurements and CV statistics.
schema.Publication	The source document a record's values come from.
schema.ExperimentReference	Reference (wildtype/control) context for an experiment.
schema.Experiment	Base experiment pairing a genotype and environment with a phenotype.
schema.FitnessExperimentReference	Reference context for a fitness experiment.
schema.FitnessExperiment	Experiment measuring a fitness phenotype.
schema.GeneInteractionExperimentReference	Reference context for a gene interaction experiment.
schema.GeneInteractionExperiment	Experiment measuring a gene interaction phenotype.
schema.GeneEssentialityExperimentReference	Reference context for a gene essentiality experiment.
schema.GeneEssentialityExperiment	Experiment measuring a gene essentiality phenotype.
schema.SyntheticLethalityExperimentReference	Reference context for a synthetic lethality experiment.
schema.SyntheticLethalityExperiment	Experiment measuring a synthetic lethality phenotype.
schema.SyntheticRescueExperimentReference	Reference context for a synthetic rescue experiment.
schema.SyntheticRescueExperiment	Experiment measuring a synthetic rescue phenotype.
schema.CalMorphExperimentReference	Reference context for a CalMorph experiment.
schema.CalMorphExperiment	Experiment measuring a CalMorph phenotype.
schema.MicroarrayExpressionPhenotype	Microarray expression phenotype with canonical log2 ratio convention.
schema.MicroarrayExpressionExperimentReference	Reference context for a microarray expression experiment.
schema.MicroarrayExpressionExperiment	Experiment measuring a microarray expression phenotype.
schema.RNASeqExpressionPhenotype	NGS (RNA-seq) expression phenotype: ABSOLUTE per-gene expression, not a ratio.
schema.RNASeqExpressionExperimentReference	Reference context for an RNA-seq expression experiment.
schema.RNASeqExpressionExperiment	Experiment measuring an RNA-seq (absolute TPM) expression phenotype.
schema.PseudobulkExpressionPhenotype	Pseudobulk single-cell RNA-seq expression: per-gene log2 fold-change vs WT, plus the per-genotype single-cell summary scalars that a bulk assay cannot provide.
schema.PseudobulkExpressionExperimentReference	Reference context for a pseudobulk single-cell expression experiment.
schema.PseudobulkExpressionExperiment	Experiment measuring a pseudobulk single-cell (Perturb-seq) expression phenotype.
schema.VisualScorePhenotype	Ordinal visual-inspection score as a proxy for a metabolite/product level.
schema.VisualScoreExperimentReference	Reference (control colony) context for a visual-score experiment.
schema.VisualScoreExperiment	Experiment measuring a visual-score phenotype.
schema.MetabolitePhenotype	Quantitative metabolite/product level(s), keyed by metabolite id.
schema.MetaboliteExperimentReference	Reference (control) context for a metabolite experiment.
schema.MetaboliteExperiment	Experiment measuring a metabolite phenotype.
schema.ProteinAbundancePhenotype	Quantitative protein-abundance profile keyed by protein (systematic ORF).
schema.ProteinAbundanceExperimentReference	Reference (control) context for a protein-abundance experiment.
schema.ProteinAbundanceExperiment	Experiment measuring a protein-abundance phenotype.
schema.AssayType	HOW an environment-response readout was physically measured (experimental design).
schema.MeasurementType	What an environment-response readout number IS, so heterogeneous chemogenomic scores are never silently compared.
schema.ResponseCategory	The controlled vocabulary a qualitative environment-response call resolves to.
schema.EnvironmentResponsePhenotype	A strain's fitness/growth RESPONSE to an environmental perturbation.
schema.EnvironmentResponseExperimentReference	Reference (control) context for an environment-response experiment.
schema.EnvironmentResponseExperiment	Experiment measuring a strain's response to an environmental perturbation.
schema.StrainEnvironmentResponseExperimentReference	Reference for a strain-resolved environment response (typed background).
schema.StrainEnvironmentResponseExperiment	A strain's response to an environmental perturbation, with a typed background.
schema.HaplotypeBlock	One run of consecutive markers called to the same parent on one chromosome.
schema.SegregantParent	One parent of a biparental cross, pinned to a sha256-anchored assembly.
schema.SegregantGenotype	Strain-level haplotype mosaic of one haploid segregant from a two-parent cross.
schema.SegregantGrowthExperimentReference	Reference (control) context for a segregant growth experiment.
schema.SegregantGrowthExperiment	Growth of a haploid segregant (haplotype-mosaic genotype) in one environment.

Functions

schema.heterozygous_deletion_functional_copies	Functional copies left at the deleted locus, respecting the background.
schema.derive_se	Derive the standard error of the mean from a reported uncertainty + its kind.

strain_background

Shared constructors for typed strain backgrounds (issue #507).

Classes

strain_background.AlleleSpec	The locus and edit kind a standard allele designation denotes.
strain_background.ByGenotype	A standard BY strain as mating type, ploidy and allele -> zygosity.

Functions

strain_background.pending_source_review	A deferred_pending_source_review gap on field naming its resolver.
strain_background.standard_allele	A BackgroundAllele for a STANDARD_ALLELES designation.
strain_background.standard_background	A EukaryoticStrainBackground for BY4741 / BY4742 / BY4743.

synthetic_lethality_to_fitness_conversion

Convert synthetic-lethality experiments into equivalent fitness experiments.

Classes

synthetic_lethality_to_fitness_conversion.SyntheticLethalityToFitnessConverter

Converter turning synthetic-lethality records into fitness records.

Functions

synthetic_lethality_to_fitness_conversion.synthetic_lethality_to_fitness_experiment	Map a synthetic-lethal experiment to a fitness-0 experiment, else drop it.
synthetic_lethality_to_fitness_conversion.synthetic_lethality_to_fitness_reference	Map a synthetic-lethality reference to a fitness-1 reference experiment.