cellink.tl.external.genescores_to_abc_road_bedgraph

cellink.tl.external.genescores_to_abc_road_bedgraph(genescores, roadmap_links, abc_links, *, roadmap_gene_col=None, roadmap_activity_col=None, abc_gene_col=None, abc_activity_col=None, chr_col='chr', start_col='start', end_col='end', use_bedtools_for_merge=True)

Convert gene scores to ABC_Road bedgraphs (Roadmap ∪ ABC strategy).

This is the primary sc-linker S2G strategy. For each program (column in genescores), creates a bedgraph where each interval’s score is the weighted sum of gene scores linked to that interval:

score(interval) = gene_score C link_activity

The Roadmap and ABC bedgraphs are merged (union) with scores summed where they overlap.

Parameters:

• genescores (DataFrame) – DataFrame (genes C programs) with probabilistic scores in [0, 1]. Index must be HGNC gene symbols.

• roadmap_links (DataFrame) – Roadmap enhancer-gene links DataFrame (from load_roadmap_links).

• abc_links (DataFrame) – ABC model enhancer-gene links DataFrame (from load_abc_links).

• roadmap_gene_col (str | None (default: None)) – Column in roadmap_links with target gene names.

• roadmap_activity_col (str | None (default: None)) – Activity/weight column in roadmap_links. If None, uses weight=1.

• abc_gene_col (str | None (default: None)) – Column in abc_links with target gene names.

• abc_activity_col (str | None (default: None)) – Activity/weight column in abc_links. If None, uses weight=1.

• chr_col (str (default: 'chr')) – Genomic coordinate columns in link DataFrames.

• start_col (str (default: 'start')) – Genomic coordinate columns in link DataFrames.

• end_col (str (default: 'end')) – Genomic coordinate columns in link DataFrames.

• use_bedtools_for_merge (bool (default: True)) – If True and bedtools is available, uses bedtools merge to merge overlapping intervals (matches original pipeline exactly). If False, uses a pure-Python merge (faster but may differ slightly).

Return type:

dict[str, DataFrame]

Returns:

dict Program name → bedgraph DataFrame (chr, start, end, score).