HOROSCOPE: Decoding human centromere architecture from short reads using k-mer signatures

Area of Science:

• Genomics
• Human Genetics
• Bioinformatics

Background:

• Centromeres are crucial for genome integrity, directing kinetochore formation and chromosome segregation.
• The repetitive nature of centromeric DNA, composed of alpha-satellite higher-order repeats (HORs), has historically limited their study using short-read sequencing.
• Understanding centromere roles in human disease is hindered by challenges in analyzing their complex architecture.

Purpose of the Study:

• To develop a computational framework, HORoSCOPE (Higher-Order Repeat organization and Size of Centromeres using Oligonucleotide Profiles for Estimation), for inferring centromere structure and length from short-read sequencing data.
• To enable large-scale population genomics of centromeric architectures.
• To investigate the relationship between centromeric alterations and human diseases, particularly cancer.

Main Methods:

• HORoSCOPE utilizes k-mer-based analysis of short-read data to infer centromere organization and size.
• A reference atlas of 11,836 human centromeres from telomere-to-telomere assembled haplotypes was created to derive chromosome-specific k-mer signatures.
• The framework leverages diagnostic k-mers to classify centromere architectures and lengths.

Main Results:

• HORoSCOPE accurately classifies chromosome-specific centromere architectures with 99.3% precision and 99.5% recall.
• Population-scale analysis of 4,029 human samples revealed continental haplotype structures and identified African-enriched rare centromeric architectures.
• Analysis of 1,359 cancer genomes linked higher-order repeat (HOR) truncation to copy-number alterations and showed chromosomal rearrangement locations depend on the centromere dip region (CDR).

Conclusions:

• HORoSCOPE provides a powerful computational framework for large-scale centromere genomics using readily available short-read data.
• The study uncovered significant population-level variation in centromeric architectures and identified potential disease associations.
• This work opens new avenues for studying centromere function and dysfunction in human health and disease.