cellink.tl.external.compute_ld_scores_for_sclinker

cellink.tl.external.compute_ld_scores_for_sclinker(annotation_prefixes, bim_prefix, ld_scores_dir, *, hapmap3_snps_prefix=None, hapmap3_snps_file=None, chromosomes=None, n_jobs=4, runner=None)

Compute LD scores for all sc-linker annotations.

Calls the existing cellink compute_ld_scores_with_annotations_from_bimfile for each (program, strategy, chromosome) combination.

Chromosome jobs are dispatched in parallel using a thread pool, each job is an independent Singularity/subprocess call so threads work well.

Parameters:

• annotation_prefixes (dict[str, dict[str, str]]) – Output of genescores_to_annotations.

• bim_prefix (str) – Prefix for per-chromosome BIM/PLINK files.

• ld_scores_dir (str | Path) – Directory to write LD score files.

• hapmap3_snps_file (str | Path | None (default: None)) – Path to a single HapMap3 SNP list file (e.g. hm3_no_MHC.list.txt from get_1000genomes_hapmap3()). Used as --print-snps for every chromosome. Takes precedence over hapmap3_snps_prefix.

• hapmap3_snps_prefix (str | None (default: None)) – Prefix for per-chromosome HapMap3 SNP files (e.g. "refs/hm."). Use hapmap3_snps_file instead when you have a single combined file.

• chromosomes (list[int] | None (default: None)) – Chromosomes to process. For a quick tutorial run use [21, 22].

• n_jobs (int (default: 4)) – Number of chromosomes to process in parallel. Each job is one Singularity subprocess call; n_jobs=4 is a safe default on an HPC node. Set to 1 to disable parallelism.

• runner (default: None) – cellink LDSCRunner. If None, uses global runner.

Return type:

dict[str, dict[str, str]]

Returns:

dict {program: {strategy: ld_score_prefix}} where ld_score_prefix ends in . (chromosome appended by LDSC).