Automated annotation of human centromeres with HORmon

Abstract

Recent advances in long-read sequencing opened a possibility to address the long-standing questions about the architecture and evolution of human centromeres. They also emphasized the need for centromere annotation (partitioning human centromeres into monomers and higher-order repeats (HORs)). Even though there was a half-century-long series of semi-manual studies of centromere architecture, a rigorous centromere annotation algorithm is still lacking. Moreover, an automated centromere annotation is a prerequisite for studies of genetic diseases associated with centromeres, and evolutionary studies of centromeres across multiple species. Although the monomer decomposition (transforming a centromere into a monocentromere written in the monomer alphabet) and the HOR decomposition (representing a monocentromere in the alphabet of HORs) are currently viewed as two separate problems, we demonstrate that they should be integrated into a single framework in such a way that HOR (monomer) inference affects monomer (HOR) inference. We thus developed the HORmon algorithm that integrates the monomer/HOR inference and automatically generates the human monomers/HORs that are largely consistent with the previous semi-manual inference.

TandemAligner: a new parameter-free framework for fast sequence alignment

“…The blue path in Figure Alignment, left shows the rare-alignment path between cenX 1 and cenX 2 constructed by TandemAlignment. This path illustrates a very different and complex evolutionary scenario with only 1,954,622 matched positions and 2,335,879 insertions and deletions -surprisingly, the orange and blue paths in To illustrate limitations of the standard alignment using a simpler example, we generated the HOR decomposition of cenX 1 (Kunyavskaya et al 2022) and extracted ten HOR-blocks at coordinates cenX 1 :1,222,223-1,242,795 resulting in a sequence of length 20,572 bp referred to as a Template. We remove the third (eighth) HOR block at coordinates Template:4,172 (14,461), and refer to the resulting sequence as Template -3 (Template -8 ).…”

“…To generate realistic simulated centromeres, we partitioned cenX 1 into HORs (Kunyavskaya et al 2022) and introduced N HOR-indel-runs. For the length of introduced HOR-indels, we use 1-based Poisson distribution with the mean estimated from Figure IndelHistogram, left and equal to 1.66.…”

“…Since tandem repeats rapidly accumulate changes in the copy numbers, they greatly differ across the human population and their expansion is known to be the causal factor of various diseases (Bakhtiari et al 2021;Park et al 2022). For example, alpha satellite arrays that host centromeres are formed by extra-long tandem repeats (ETRs) representing some of the most rapidly evolving regions of the human genome (Dvorkina et al 2021;Altemose et al 2022;Kunyavskaya et al 2022). Comparison of ETRs across the human population is a prerequisite for understanding their evolution and association with cancer and infertility (Schueler et al 2001;Alkan et al 2007;Enukashvily et al 2007;Shepelev et al 2009;Ting et al 2011;Nagaoka, Hassold, and Hunt 2012;Melters et al 2013;Giunta and Funabiki 2017;Black and Giunta 2018;Smurova and De Wulf 2018;Miga 2019;Miga and Alexandrov 2021;Altemose et al 2022) .…”

HiCAT: a tool for automatic annotation of centromere structure

“…To address this question, Dvorkina et al proposed the first automatic centromere annotation tool, CentromereArchitect [ 10 ], which was based on StringDecomposer (SD) [ 4 ], an algorithm for detecting sequence blocks by taking monomer templates to decompose centromere DNA sequences. In CentromereArchitect, monomer inference and HOR detection were considered two separate problems without interconnections, which often led to biologically inadequate annotation [ 14 ]. The authors next proposed HORmon [ 14 ] based on the centromere evolution postulate (CE postulate, where each monomer appears only once in the HOR unit) to address the lack of interconnection issue in CentromereArchitect.…”

“…In CentromereArchitect, monomer inference and HOR detection were considered two separate problems without interconnections, which often led to biologically inadequate annotation [ 14 ]. The authors next proposed HORmon [ 14 ] based on the centromere evolution postulate (CE postulate, where each monomer appears only once in the HOR unit) to address the lack of interconnection issue in CentromereArchitect. HORmon first constructs a de Bruijn graph based on monomers inferred from CentromereArchitect and then refines the monomers by considering positional similarity to amend the graph as a single cycle (referred to as the detected HOR) to comply with the CE postulate.…”