Skip to content

CHANGELOG.md

Latest commit

 

History

History
74 lines (58 loc) · 3.7 KB

CHANGELOG.md

File metadata and controls

74 lines (58 loc) · 3.7 KB

Changelog

All notable changes to md-python are documented in this file.

The format follows Keep a Changelog and the project loosely adheres to Semantic Versioning.

[0.3.4]

  • Updated client.entities.mappings.peptide_to_protein_same_dataset and client.entities.mappings.protein_to_protein_via_peptides to accept a list of datasets instead of a single dataset.

[0.3.3]

  • client.entities.mappings.protein_to_protein_via_peptides map protein groups to protein groups through their shared peptides across datasets.

[0.3.2]

  • client.entities.mappings.peptide_to_protein_same_dataset map peptides to their protein groups within a single dataset.

[0.3.1]

  • client.entities.mappings.protein_to_peptide_same_dataset map protein groups to their peptides within a single dataset.

[0.3.0]

Added

  • NormalisationImputationDataset now exposes the full surface of the MD Converter normalisation_and_imputation flow:
    • Filtration for entity_type="protein" via filtration_method="by missing values" (previously blocked).
    • knn_tn imputation with knn_tn_k (1–10, default 5) and knn_tn_distance (truncation | correlation, default truncation).
    • mindet imputation parameter q (0–0.5, default 0.01).
    • Batch-correction sub-technique selector batch_correction_technique (limma remove batch effect | combat | combat seq — combat seq is gene-only) with the matching parameter blocks (batch_variables for limma, batch_variable_combat
      • mean_only + reference_batch_combat for ComBat, batch_variable_combat only for ComBat-Seq, plus design_variables and experiment_design).
    • include_imputed_values (default False) on median / quantile / sum / batch correction.
    • median_normalisation_centre_at_zero (default True) on median normalisation.
    • Shared filter block for by missing values and by minimum abundance: filter_valid_values_criteria (percentage | count), filter_threshold_proportion, filter_threshold_count, filter_valid_values_logic, filter_based_on_condition.
    • NormalisationImputationDataset.filter_only(...) classmethod for filtration-only jobs (normalisation=skip, imputation=skip, output remains INTENSITY).
    • NormalisationImputationDataset.help() classmethod with an in-process method-and-parameter reference.

Changed

  • Wire-format strings emitted by NormalisationImputationDataset now use the MD Converter canonical (spaced) form: "batch correction", "by missing values", "by ptm localization probability", "by minimum abundance", "limma remove batch effect", "combat seq". Underscored variants ("batch_correction", "minimum_abundance", "ptm_localization_probability", "by_missing_values", "limma_remove_batch_effect", "combat_seq") are still accepted on input and normalised to canonical on output.
  • NormalisationImputationDataset constructor moved from the legacy nested normalisation_methods / imputation_methods dicts to flat typed kwargs (normalisation_method, imputation_method, entity_type, filtration_method, plus method-specific kwargs). The extra_params escape hatch is retained and is merged last, so caller-supplied keys override typed defaults.

Notes

  • The output dataset type for the NI flow is INTENSITY — including the filter-only pattern via filter_only(...). An upload may therefore have multiple INTENSITY datasets after running NI; find_initial_dataset does not yet disambiguate this and will raise. A follow-up release will refine this.

[0.2.4]

  • Prior releases — see git history.