mecfs_bio.assets.gwas.multi_trait.lcv.mi_lcv_analysis
Apply LCV to a list of a candidate upstream causal traits against the downstream trait Myocardial Infarction.
The goal here is mainly to make sure my LCV implementation is correct by testing for well known causal factors like LDL.
Attributes:
MI_LCV_TASK_GROUP
module-attribute
MI_LCV_TASK_GROUP = lcv_generate(
base_name="mi_analysis",
upstream_traits=[
LCVSourceTraitInfo(
name="LDL",
df_task=parquet_file_task,
pipe=CompositePipe(
[
FilterRowsByMinInCol(
1e-15, col=GWASLAB_SE_COL
)
]
),
),
LCVSourceTraitInfo(
name="Triglycerides",
df_task=parquet_file_task,
pipe=CompositePipe(
[
FilterRowsByMinInCol(
1e-15, col=GWASLAB_SE_COL
)
]
),
),
LCVSourceTraitInfo(
name="CRP",
df_task=parquet_file_task,
pipe=CompositePipe(
[
FilterRowsByMinInCol(
1e-15, col=GWASLAB_SE_COL
)
]
),
),
LCVSourceTraitInfo(
name="Educational_Attainment",
df_task=parquet_file_task,
pipe=CompositePipe(
[
FilterRowsByMinInCol(
1e-15, col=GWASLAB_SE_COL
)
]
),
),
],
downstream_traits=[
LCVSourceTraitInfo(
name="MI",
df_task=parquet_file_task,
pipe=CompositePipe(
[
ComputeBetaIfNeededPipe(),
ComputeSEPipe(),
FilterRowsByMinInCol(
1e-15, col=GWASLAB_SE_COL
),
ToPolarsPipe(),
]
),
)
],
consolidated_ld_scores_task=THOUSAND_GENOME_EUR_LD_REFERENCE_DATA_V1_CONSOLIDATE,
config=LCVConfig(chisq_exclude_factor_threshold=50),
)