Skip to content

mecfs_bio.asset_generator.protein_coloc_asset_generator

Asset generator for colocalizing one trait GWAS with the proteins of one proteomics GWAS (UKB-PPP plasma, Western et al. CSF).

Builds, for one GRCh38-harmonized trait: the TraitProteinColocTask (raw results and aligned windows) and the docs-facing display table of the pairs with PP.H4 above a threshold. Stacked locus plots are not built per run: generate_coloc_locus_plot_task makes them for hand-picked pairs of a run. The proteomics source supplies the hit table, the dense window fetcher and the gene coordinates; the GRCh38 genome, allele-frequency panel and gene track are shared by every source.

Classes:

Functions:

Attributes:

ProteomicsSourceInfix module-attribute 

ProteomicsSourceInfix = Literal['ppp', 'csf']

ColocStackPlotGroup

plot_task: the per-locus stacked plots (PNG, SVG and an index table in one directory). svg_task: just the SVG figures, copied out of plot_task's directory. png_task: just the PNG figures, copied out of plot_task's directory.

Methods:

Attributes:

plot_task instance-attribute 

plot_task: ColocLocusStackPlotTask

png_task instance-attribute 

png_task: CopyFilesWithSuffixFromDirectoryTask

svg_task instance-attribute 

svg_task: CopyFilesWithSuffixFromDirectoryTask

terminal_tasks 

terminal_tasks() -> list[Task]
Source code in mecfs_bio/asset_generator/protein_coloc_asset_generator.py 
154
155
def terminal_tasks(self) -> list[Task]:
    return [self.svg_task, self.png_task]

ColocTaskGroup

Methods:

Attributes:

coloc_task instance-attribute 

coloc_task: TraitProteinColocTask

display_task instance-attribute 

display_task: ColocDisplayTableTask

terminal_tasks 

terminal_tasks() -> list[Task]
Source code in mecfs_bio/asset_generator/protein_coloc_asset_generator.py 
91
92
def terminal_tasks(self) -> list[Task]:
    return [self.display_task]

ProteomicsColocSource

asset_infix: names the source in the asset ids ({base}_{infix}_coloc, ...). hit_table_task: the source's protein hit-segment table. gene_coords_task, gene_coords_reader: the GRCh38 gene-coordinate table and its reader. fetcher: dense window source of the proteomics GWAS.

Attributes:

asset_infix instance-attribute 

asset_infix: ProteomicsSourceInfix

fetcher instance-attribute 

fetcher: DenseWindowFetcher

gene_coords_reader instance-attribute 

gene_coords_reader: GeneCoordsReader

gene_coords_task instance-attribute 

gene_coords_task: Task

hit_table_task instance-attribute 

hit_table_task: ProteinHitTableTask

generate_coloc_locus_plot_task 

generate_coloc_locus_plot_task(
    asset_id: str,
    group: ColocTaskGroup,
    pairs: tuple[ColocPair, ...],
) -> ColocLocusPlotTask

Stacked locus plots of chosen tested pairs of one coloc run.

Source code in mecfs_bio/asset_generator/protein_coloc_asset_generator.py 
130
131
132
133
134
135
136
137
138
139
def generate_coloc_locus_plot_task(
    asset_id: str, group: ColocTaskGroup, pairs: tuple[ColocPair, ...]
) -> ColocLocusPlotTask:
    """Stacked locus plots of chosen tested pairs of one coloc run."""
    return ColocLocusPlotTask.create(
        asset_id=asset_id,
        coloc_task=group.coloc_task,
        gene_info_task=MAGMA_ENTREZ_GENE_LOCATION_REFERENCE_DATA_BUILD_38_EXTRACTED,
        pairs=pairs,
    )

generate_coloc_locus_stack_plot_task 

generate_coloc_locus_stack_plot_task(
    asset_id: str,
    group: ColocTaskGroup,
    max_loci: int = 10,
    trait_name: str | None = None,
) -> ColocStackPlotGroup

Per-locus stacked trait and protein Manhattan plots of the top display-table loci.

trait_name, if given, labels the trait row in place of "trait".

The SVG and PNG figures are also copied out into their own directories (asset ids with _svg and _png appended).

Source code in mecfs_bio/asset_generator/protein_coloc_asset_generator.py 
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
def generate_coloc_locus_stack_plot_task(
    asset_id: str,
    group: ColocTaskGroup,
    max_loci: int = 10,
    trait_name: str | None = None,
) -> ColocStackPlotGroup:
    """Per-locus stacked trait and protein Manhattan plots of the top display-table loci.

    trait_name, if given, labels the trait row in place of "trait".

    The SVG and PNG figures are also copied out into their own directories (asset ids with
    _svg and _png appended).
    """
    plot_task = ColocLocusStackPlotTask.create(
        asset_id=asset_id,
        coloc_task=group.coloc_task,
        trait_task=group.coloc_task.trait_task,
        display_task=group.display_task,
        gene_info_task=MAGMA_ENTREZ_GENE_LOCATION_REFERENCE_DATA_BUILD_38_EXTRACTED,
        genetic_map_task=GENETIC_MAP_HG38,
        max_loci=max_loci,
        trait_name=trait_name,
    )
    return ColocStackPlotGroup(
        plot_task=plot_task,
        svg_task=CopyFilesWithSuffixFromDirectoryTask.create(
            asset_id=f"{asset_id}_svg", source_directory_task=plot_task, suffix=".svg"
        ),
        png_task=CopyFilesWithSuffixFromDirectoryTask.create(
            asset_id=f"{asset_id}_png", source_directory_task=plot_task, suffix=".png"
        ),
    )

generate_coloc_protein_info_task 

generate_coloc_protein_info_task(
    asset_id: str, group: ColocTaskGroup, max_loci: int = 10
) -> ColocProteinInfoTask

UniProt and Ensembl annotations of the colocalizing proteins at the top display-table loci.

Only UKB-PPP runs are supported: the UniProt accessions of each protein are read from the Sun et al. 2023 ST3 sheet, the run's gene-coordinate table.

Source code in mecfs_bio/asset_generator/protein_coloc_asset_generator.py 
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
def generate_coloc_protein_info_task(
    asset_id: str, group: ColocTaskGroup, max_loci: int = 10
) -> ColocProteinInfoTask:
    """UniProt and Ensembl annotations of the colocalizing proteins at the top display-table loci.

    Only UKB-PPP runs are supported: the UniProt accessions of each protein are read from the
    Sun et al. 2023 ST3 sheet, the run's gene-coordinate table.
    """
    coloc_task = group.coloc_task
    assert isinstance(coloc_task.gene_coords_reader, St3GeneCoordsReader), (
        "generate_coloc_protein_info_task supports UKB-PPP runs only, got gene coords reader "
        f"{type(coloc_task.gene_coords_reader).__name__}"
    )
    return ColocProteinInfoTask.create(
        asset_id=asset_id,
        display_task=group.display_task,
        protein_table_task=coloc_task.gene_coords_task,
        protein_accessions_reader=St3ProteinAccessionsReader(),
        uniprot_task=UNIPROT_LOOKUP,
        uniprot_genes_task=HGNC_UNIPROT_GENES,
        max_loci=max_loci,
    )

generate_protein_coloc_tasks 

generate_protein_coloc_tasks(
    base_name: str,
    source: ProteomicsColocSource,
    trait_task: Task,
    trait_model: TraitModel,
    config: ColocRunConfig,
    h4_threshold: float = 0.8,
) -> ColocTaskGroup

base_name: prefix of the asset ids. trait_task: the trait GWAS, genome-reference harmonized to GRCh38. trait_model: quantitative or case-control, as coloc.abf needs.

Source code in mecfs_bio/asset_generator/protein_coloc_asset_generator.py 
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
def generate_protein_coloc_tasks(
    base_name: str,
    source: ProteomicsColocSource,
    trait_task: Task,
    trait_model: TraitModel,
    config: ColocRunConfig,
    h4_threshold: float = 0.8,
) -> ColocTaskGroup:
    """
    base_name: prefix of the asset ids.
    trait_task: the trait GWAS, genome-reference harmonized to GRCh38.
    trait_model: quantitative or case-control, as coloc.abf needs.
    """
    prefix = f"{base_name}_{source.asset_infix}"
    coloc_task = TraitProteinColocTask.create(
        asset_id=f"{prefix}_coloc",
        trait_task=trait_task,
        trait_model=trait_model,
        hit_table_task=source.hit_table_task,
        gene_coords_task=source.gene_coords_task,
        gene_coords_reader=source.gene_coords_reader,
        fasta_task=UCSC_HG38_INDEXED_FASTA,
        panel_task=GNOMAD_V4_1_GENOMES_HG38_ALLELE_FREQUENCIES,
        panel_ancestry="nfe",
        fetcher=source.fetcher,
        config=config,
    )
    display_task = ColocDisplayTableTask.create(
        asset_id=f"{prefix}_coloc_display",
        coloc_task=coloc_task,
        h4_threshold=h4_threshold,
    )
    return ColocTaskGroup(coloc_task=coloc_task, display_task=display_task)