torchcell.datasets.scerevisiae.kuzmin2018.DmfKuzmin2018Dataset

class DmfKuzmin2018Dataset(root: str = 'data/torchcell/dmf_kuzmin2018', subset_n: int | None = None, io_workers: int = 0, transform: Callable[[...], Any] | None = None, pre_transform: Callable[[...], Any] | None = None, **kwargs: Any)

Bases: ExperimentDataset

Double-mutant fitness experiments from Kuzmin 2018.

Two record kinds (record_kind): the digenic query x array crosses, and the double-mutant QUERY strains of the trigenic screens, which the raw table also measures (“Query single/double mutant fitness”), one record per distinct strain.

organism: ClassVar[Organism] = 'scerevisiae'

each loader declares its own, the build’s organism switch reads it off the class (dataset_organism), and a loader that forgot it is refused by name instead of inheriting an organism it may not have.

Type:

The host organism every record describes. Annotated here WITHOUT a value

property experiment_class: type[Experiment]

Return the experiment model class for this dataset.

property reference_class: type[ExperimentReference]

Return the experiment-reference model class for this dataset.

property raw_file_names: list[str]

Data S1 (from the hosted zip) and Additional Data S4 (from the raw mirror).

download() → None

Fetch Data S1 from the hosted zip unless it is in raw/; link S4 in.

Data S1 is fetched only when absent so a store whose Data S1 is already in raw/ gains Additional Data S4 without a network call.

process() → None

Preprocess the raw table and write experiment records to the LMDB store.

preprocess_raw(df: DataFrame, df_s4: DataFrame) → DataFrame

Clean and reshape the raw Kuzmin table into per-genotype rows.

The query-strain rows gain query_strain_sd from Additional Data S4 (_attach_query_strain_sd).

static create_experiment(dataset_name: str, row: Series, phenotype_reference_std: float) → tuple[FitnessExperiment, FitnessExperimentReference, Publication]

Build the experiment, reference, and publication objects from a data row.