# ensembl115_human_paralogs.tsv.gz -- deposited 2026-08-03
#
# Human within-species paralogs with protein percent identity, as returned by biomaRt for the
# genes classified in this study. Deposited because homology is NOT derivable from the GTFs in
# this record: a GTF states where exons are, never that gene A is a paralog of gene B at 82%.
# 05u_paralog_annotation.R used to fetch this live, which made the whole chain need internet --
# and this table defines the paralog screen behind the train/test split, so a reader who cannot
# fetch it loses the leakage screen rather than a figure.
#
# biomaRt host     : https://sep2025.archive.ensembl.org:443/biomart/martservice
# dataset          : hsapiens_gene_ensembl, version 115
# attributes       : ensembl_gene_id, hsapiens_paralog_ensembl_gene, hsapiens_paralog_perc_id
# queried          : 2026-07-29 20:30:05
# rows             : 131280
# genes queried    : 14303
