torchcell.datamodels.schema.StrainBackground

class StrainBackground(*, provenance_gaps: list[~torchcell.verification.sourced.ProvenanceGap] = <factory>, name: str, reference_strain: ~typing.Literal['S288C', 'CENPK113-7D', 'MG1655', 'BW25113', 'KT2440', 'REL606', 'W3110'], assembly_set: ~typing.Literal['ecoli_K12_MG1655_ASM584v2', 'ecoli_K12_BW25113_ASM75055v1', 'pputida_KT2440_ASM756v2', 'ecoli_B_REL606_ASM1798v1', 'ecoli_K12_W3110_ASM1024v1', 'sgd_S288C_R64-4-1_20230830', 'scerevisiae_CENPK113-7D_ASM257140v2'], parents: list[str] | None = None, construction: str | None = None, genotype_statement: str | None = None, alleles: list[~torchcell.datamodels.schema.BackgroundAllele] = <factory>, provenance: list[~torchcell.verification.sourced.SourcedValue] | None = None)[source]

Bases: ProvenanceGapMixin

The genome content a strain carries beyond its reference genome.

ONE class for every host (P5, ontology-unify 2026-10-15), holding what both families state: the strain label, the sequenced strain the alleles are edits against, the deposited bytes that strain means, the parents, the construction, the source’s verbatim genotype string and the typed alleles. The retired BacterialStrainBackground differed from this by exactly the three eukaryote facts, which EukaryoticStrainBackground carries, and the schema.py comment that chose the duplication chose it on cost.

name is the strain or lineage label the records join on (BY4741, BW25113, or a descriptive name for an SGA progeny pool); it must equal StrainReferenceGenome.strain. parents are the strains it was made from, verbatim (["BY4741", "BY4742"] for BY4743; the SGA query and array for a progeny pool). construction is a one-line method statement (how the background was made, e.g. an SGA cross and its selections). genotype_statement is the genotype exactly as the source writes it, kept beside the typed alleles so nothing the source said is lost: BW25113’s lesions come from its own deposited GenBank source feature rather than from a paper, and a yeast strain table states the same kind of string.

reference_strain is the sequenced strain every allele is an edit against and assembly_set is the deposited set that strain means; the two must agree (REFERENCE_STRAIN_ASSEMBLY_SETS). The pin is REQUIRED with no default and no typed gap, which is the one place this field’s contract differs from ReferenceGenome.assembly_set: that carrier’s companion field is the free string strain and may name a strain the tier has not deposited, so it needs the gap, while reference_strain here is a closed vocabulary with one member per deposited set, so a background that can name its reference has already named a deposited set and a gap branch would be unreachable. A background whose reference is not deposited is not representable, deliberately: its edit list against another reference would be thousands of variants.

provenance is set or gapped (a typed absence, never a silent None). A single-copy genome carries one allele entry per locus and states no copy count, so every allele’s zygosity is None here; EukaryoticStrainBackground is where a copy count is a fact and is required.

alleles_at(systematic_gene_name: str) → list[BackgroundAllele][source]: The background’s allele entries at one locus (empty = the reference allele).

functional_copies(systematic_gene_name: str) → int[source]

Functional copies of a gene in the UNPERTURBED background: 1 or 0.

One copy, so the answer is 1 unless the background carries a non-functional allele at the locus. EukaryoticStrainBackground overrides this with the ploidy-aware count.

property is_fully_sourced: bool: True when the background and every allele carry no declared gap.

model_config: ClassVar[ConfigDict] = {'extra': 'forbid', 'frozen': True}: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].