SNV effects on uninterrupted repeats
Does an HG002 SNV lengthen or shorten the longest pure repeat run? Here, A contains variants and B contains repeat intervals, motifs, and reference sequences. The Python callback substitutes each ALT and reports the change.
Python callback
This excerpt uses effect() from the
full script.
That helper compares the longest consecutive run of complete motif copies,
allowing all motif rotations and overlapping start positions.
def bedder_repeat_run(fragment) -> str:
"""Repeat ID|motif|reference pure copies|alternate pure copies|delta; single-SNV effect."""
variant = fragment.a.vcf()
if variant is None:
raise ValueError('requires VCF query records')
if len(variant.ALT) != 1:
raise ValueError('requires biallelic input')
if len(fragment.b) != 1:
raise ValueError('requires one repeat per record: use --a-piece whole --b-piece whole')
repeat = fragment.b[0].bed()
if repeat is None:
raise ValueError('requires BED repeat records')
motif, sequence = repeat.other_fields()
if len(sequence) != repeat.stop - repeat.start:
raise ValueError('repeat sequence length does not match BED interval')
before, after, delta = effect(sequence, motif, variant.pos - repeat.start,
variant.REF, variant.ALT[0])
return f'{repeat.name}|{motif}|{before}|{after}|{delta}'
Run it
From a bedder-rs checkout on the manuscript branch, using the full script and
bundled inputs:
example=manuscript/repeat-interruptions
bedder intersect -a "$example/data/variants.vcf" \
-b "$example/data/repeats.bed" -g "$example/data/genome.tsv" \
--a-piece whole --b-piece whole \
--python "$example/repeat_run_effect.py" -c py:repeat_run \
-o repeat-effects.vcf
Each output record describes one variant–repeat pair. For example, chr19:16259432 T>C changes the longest uninterrupted AC run from 13 to 26 copies, without inserting bases. The frozen example validates 561 pairs from 551 HG002 SNVs.
This is one SNV applied to the reference sequence, bounded by the repeat interval—not a reconstructed HG002 haplotype or a pathogenicity prediction.