From IBD tracts to runs of homozygosity: a unified coalescent framework including selection
summary
The gist
Identity by descent (IBD) tracts and runs of homozygosity (ROH) represent the theoretical and observable sides of chromosomal autozygosity The formal relationship between their length distributions…
In short
The paper unifies Identity by Descent (IBD) tracts and Runs of Homozygosity (ROH) using a single coalescent model. This framework relates their length distributions to infer demographic parameters like effective population size ($N_e$) and detect selection signatures, providing a formal link between theoretical constructs and observable genomic data.
Key concepts
- Identity by Descent (IBD) Tracts
- These are theoretical segments of the genome shared by two haplotypes descended from a single ancestor without internal recombination. They are bounded by recombination breakpoints that cannot be directly sequenced, making them a theoretical construct used to model autozygosity.
- Runs of Homozygosity (ROH)
- ROHs are observable stretches in an individual where all genetic markers within that region are homozygous. Their lengths provide empirical evidence about the time since two copies of a chromosome became identical, reflecting coalescence history.
- Coalescent Framework
- This is a mathematical model based on random sexual reproduction to track how genetic lineages merge over time. It provides a unified way to analyze both theoretical IBD tracts and observable ROH lengths within one consistent mathematical structure.
- Effective Population Size ($N_e$)
- $N_e$ estimates the size of an idealized population that would experience the same amount of genetic drift as a real-world population. The framework uses observed IBD tract lengths to calculate $N_e$, allowing researchers to infer demographic history and local population sizes.
Terminology used across episodes
This episode discusses
- From IBD tracts to runs of homozygosity: a unified coalescent framework including selection · Paper Radio
The paper
From IBD tracts to runs of homozygosity: a unified coalescent framework including selection · Read on arXiv
Enrique Santiago
Departamento de Biologia Funcional · Facultad de Biologia · Universidad de Oviedo
Identity-by-descent (IBD) tracts and runs of homozygosity (ROH) represent the underlying and observable aspects of chromosomal autozygosity, but no formal relationship between them had been established. This study develops a coalescent framework that connects these concepts, providing tools for detecting selection signatures and extending the temporal range of effective population size (Ne) inference. Closed-form probability functions are derived for IBD tract lengths and mapped onto observable ROHs by modelling the displacement of autozygosity boundaries from true recombination breakpoints to the nearest heterozygous marker. This mapping explicitly accounts for mutation, gene conversion, marker density, and heterozygosity. Selection effects are incorporated analytically, yielding computational tools to predict tract length distributions under background selection and partial or complete selective sweeps. Background selection shifts IBD tract-length distributions toward longer tracts, an effect better explained by a generation-dependent Ne reduction below its neutral value than by a single constant Ne. Selective sweeps generate directional asymmetry in tract-length distributions at neutral sites linked with the selected locus, with its direction indicating on which side of the neutral site the selected locus resides. This asymmetry decays rapidly after the sweep, whereas the local Ne reduction estimated from the median tract length persists considerably longer. The coalescent framework is further extended to arbitrary time-varying population size, enabling insights into past demography across a wide temporal range.
Transcript
Introduction to the show: ident: Genomics Radio. Generated commentary on the latest computational biology and genomics papers.
Ines: We're looking at the paper: "From IBD tracts to runs of homozygosity".
Marcus: Here's the gist of the paper. Identity by descent (IBD) tracts and runs of homozygosity (ROH) represent the theoretical and observable sides of chromosomal autozygosity The formal relationship between their length distributions has yet to be established A coalescent framework is developed here that unifies both concepts within a single analytical formalism, with applications to inferring effective population size (Ne) and detecting selection signatures Conceptual Foundation of IBD and ROH Identity by descent (IBD) tracts and runs of homozygosity (ROH) are related concepts that describe autozygosity, the state in which the two homologous copies of a genomic region are identical due to descent from a common ancestor An IBD tract is a theoretical construct representing a contiguous chromosomal segment shared by two haplotypes that descend from a single ancestral copy and have experienced no internal recombination (Thompson 2018) By contrast, an ROH is an empirically observable entity: a contiguous stretch of the genome in a diploid individual in which all typed markers are homozygous (McQuillan et al.
Ines: 2008) While IBD tracts are bounded by recombination breakpoints that cannot be located from sequence data, ROHs are clearly defined by flanking heterozygous marker sites The lengths of autozygous segments provide information about coalescence time: recently coalesced haplotypes share long autozygous segments, whereas segments derived from more ancient common ancestors have been progressively eroded by recombination Modeling Autozygosity Dynamics The general model assumes a Wright-Fisher model of 2N haploid individuals with random sexual reproduction Recombination events along each branch progressively trims the original chromosome on both sides of the focal site, defining left (L) and right (R) boundaries of the IBD tract Since recombination breakpoints cannot be observed directly from sequence data, the IBD tract cannot be precisely measured The model also incorporates de novo mutations and gene conversion events, which interrupt the identity continuity similarly to recombination Since break events of recombination, mutation, and gene conversion are each Poisson distributed, the distances between consecutive events within a generation follow an exponential distribution with a rate (1 + m) The total tract length, x = x% + x), follows an Erlang distribution with shape 2 and the same ratelambda as both exponential distributions Inference of Demographic Parameters Equations for Ne in terms of the observed IBD tract lengths serve two purposes: inferring demographic history from whole-genome marker data, and estimating local effective population size at specific genomic sites The first estimator is based on the mean length x̅ of IBD tracts containing the focal site within the genomic region of interest The second estimator is based on the proportion p@A!5;- of IBD tracts that contain both the focal site and a reference site at a distance of x Morgans These equations allow Ne to be estimated directly from the observed proportion of the genome covered by IBD tracts of a given length Incorporating Evolutionary Forces The effect of background selection on IBD tract length distributions is modeled using a generation-dependent Ne(t) The cumulative effect of selection over t generations is expressed by the equation N+(8) Selective sweeps produce time-limited distortions 1n the distribution of IBD tract lengths at sites near the selected locus The two local Ne estimators behave in a complementary manner, with the mean-based estimator detecting the sweep signal earlier and the median-based estimator persisting for a longer period Unification of Theoretical Constructs This framework formally links three areas of theory - IBD tract length distributions, observable ROH length distributions and CSH probabilities - within a single coalescent model The connection between IBD tracts and ROH was established in two steps The CSH formulation emerges as a natural special case of the framework, corresponding precisely to the probability that an IBD tract spans both members of a marker pair Validation and Application The accuracy of Equation 11 was evaluated by forward simulation of populations at mutation–drift equilibrium on twoMorgan chromosomes, in the absence of selection The predictions of Equation 7 under background selection were validated by forward simulation of a population carrying a two-Morgan chromosome When applied to human population data, local Ne minima are revealed across the genome, consistent with pervasive positive selection events The relative transient nature of the asymmetry signal suggests that the sweep may be recent Software and Tools The PDF IBD tracts BS.py and PDF IBD tracts SS.py programs can be downloaded from https://github.com/esrud The former predicts the distribution of IBD tract lengths under background selection, while the latter predicts the distribution of IBD tract lengths at any moment during a selective sweep The total ROH tract length x = x(+ xC is the sum of independent and identically distributed left and right flanking lengths This reduces to equation A2 when d → 0, in which case IBD and ROH boundaries coincide The closed-form approximation for the total length x is given by Equation A17 This reduces to equation A2 when d → 0, in which case IBD and ROH boundaries coincide The corresponding ROH length distribution to the right side of the focal site conditional on time of coalescence is P(x(;t) ≈ 2t(1 + m)∙ e"0!-(n∗+) x H / (H − d y!
Marcus: 0!∙ 2t∙ x(1 + m)− e"9M F"‡: y Final Derivations The closed-form expression for the proportion p@A!5;- of ROH tracts that include both the focal site and a reference site at distance x Morgans can be derived from Equation A15 This reduces to Equation A10 when the density of markers is infinite The corresponding ROH length distribution to the right side of the focal site conditional on time of coalescence is P(x(;t) ≈ 2t(1 + m)∙ e"0!-(n∗+) x H / (H − d y!
Ines: 0!∙ 2t∙ x(1 + m)− e"9M F"‡: y The final closed-form expression for the proportion p;<=(-,?) of the genome covered by ROH tracts within a small window of width h centered at length x is p;<=(-;?) = h∙ P(x) Discussion Summary The expected tract length under Equation 1 is half that under Equation 2, a direct consequence of length-biased sampling For very short tracts (x≪N+&"), the mean coalescence time approaches 6N+ generations, exceeding the 2N+ expectation in Kingman's (1982) pairwise coalescence model This is not inconsistent, but rather a sampling effect: short tracts are enriched for ancient coalescence events The framework provides a mean of reconstructing past trajectories from the observed tract length distribution The CSH formulation corresponds precisely to the probability that an IBD tract spans both members of a marker pair <ref:2605.
Title and authors: Ines: The paper starts by defining IBD tracts as a theoretical construct, something shared by two haplotypes descending from a single ancestor with no internal recombination, and ROHs as the observable thing—a stretch where all markers are homozygous.
Marcus: And they point out that while you can’t find the true recombination breakpoints for IBD tracts using sequence data, you can clearly define ROHs because they have flanking heterozygous markers.
Yuki: The paper then says these lengths give us information about coalescence time; recent haplotypes share longer autozygous segments because they haven't been broken up much by recombination yet.
Ines: It models the dynamics of this process assuming a Wright-Fisher model for 2N haploid individuals with random sexual reproduction, and it incorporates de novo mutations and gene conversion events that interrupt this continuity <ref:2605.03498#pg1>.
Marcus: These interruptions are modeled as Poisson distributed events, which means the distances between those breaks follow an exponential distribution with a rate of one plus m, where m is some parameter related to these interruptions.
Yuki: The total tract length then follows an Erlang distribution with shape two and the same rate lambda as those exponential distributions, which is a key mathematical link in their framework.
Ines: And crucially, they show how background selection introduces a systematic upward bias in apparent tract lengths that you can’t capture with a single constant Ne; it requires a generation-dependent Ne.
Marcus: And selective sweeps cause an asymmetric distortion of the length distribution around a neutral focal site, and the sign of that asymmetry tells us something specific about where the selected site is located.
The paper's summary: Ines: The authors propose several ways to make this framework more robust, particularly regarding how they link IBD tracts to ROHs through a coalescent model. They derive closed-form probability density functions for IBD tract lengths and then extend those directly to the observable ROH length distribution.
Marcus: This extension is key because it explicitly models how the autozygosity boundaries shift from the true recombination breakpoints out to the nearest heterozygous flanking marker sites, which links them up formally.
Yuki: They also incorporate things like finite marker density and marker heterozygosity as parameters, which is important because real genomic data isn't perfect; it’s not a clean infinite sequence.
Ines: And then they handle background selection by introducing a generation-dependent Ne that changes over time, which is necessary because selection has an effect on the tract lengths in a way that isn't constant across generations.
Marcus: Regarding selective sweeps, they model this as an asymmetric distortion around the focal site; they state that the sign of this asymmetry indicates where the selected site is located. That’s a practical application for finding genomic regions under selection.
Yuki: The paper also discusses how this framework relates to CSH probabilities—the probability that an IBD tract spans both members of a marker pair—and shows how it emerges naturally as a special case of their unified model, which is interesting for connecting these concepts.
The paper's improvements: Ines: So to wrap up, this paper gives us a single coalescent model that ties together the theoretical IBD tracts and the observable ROH lengths, while also letting us use it to estimate population size and detect selection signals.
Marcus: It’s about taking those two concepts, IBD tracts and ROHs, which we usually look at separately, and putting them into one analytical formalism to see how they interact under different evolutionary forces.
Yuki: For the wider context of species history, this means we have a more unified language to interpret autozygosity patterns across populations without having to switch between different models for tracts and ROHs.
Ines: And the results show that when you apply these equations, like Equation eleven they can be validated through forward simulations at mutation–drift equilibrium on two-Morgan chromosomes without selection, and under background selection when applied to human data, we see local Ne minima consistent with positive selection events.
Marcus: The software they released, PDF IBD tracts BS.py and PDF IBD tracts SS.py, lets us predict the distribution of tract lengths under selection or at any moment during a sweep, which is a tool for analyzing actual sequencing data from cohorts.
Yuki: It’s about providing a way to reconstruct past trajectories from the observed tract length distributions by using the mean length to infer coalescence times and demographic history.
Ines: We're looking forward to seeing how this unified framework helps us better understand the interplay between demography, selection, and autozygosity in future studies.
Marcus: It’s a solid piece of work that gives us a formal way to compare these concepts and apply them practically to infer population dynamics from genomic data.
Conclusion: Ines: So to wrap up, this paper introduces a single coalescent model that connects IBD tracts and runs of homozygosity, while also including how selection affects those tract lengths.
Marcus: Exactly. It takes those two things—the theoretical IBD tracts and the observable ROH lengths—and puts them into one analytical framework to see how they work together under different evolutionary forces.
Yuki: And what it means for us is a more unified language to interpret autozygosity patterns across populations without having to switch between different models for tracts and ROHs.
Ines: The results show that when you apply these equations, like Equation eleven they can be validated through forward simulations at mutation–drift equilibrium on two-Morgan chromosomes without selection.
Marcus: And under background selection when applied to human data, we see local Ne minima consistent with positive selection events.
Yuki: It’s about providing a way to reconstruct past trajectories from the observed tract length distributions by using the mean length to infer coalescence times and demographic history.
Ines: The software they released, PDF IBD tracts BS.py and PDF IBD tracts SS.py, lets us predict the distribution of tract lengths under selection or at any moment during a sweep, which is a tool for analyzing actual sequencing data from cohorts.
Marcus: It’s a solid piece of work that gives us a formal way to compare these concepts and apply them practically to infer population dynamics from genomic data.
Yuki: For me, the CSH formulation they derived corresponds precisely to the probability that an IBD tract spans both members of a marker pair, which is really useful for connecting this back to historical events.
Ines: We're looking forward to seeing how this unified framework helps us better understand the interplay between demography, selection, and autozygosity in future studies.
Marcus: Next time we look at a paper that tackles real-world cohort data and the statistics behind those claims.
More episodes
- 2610.33881-Sensorimotor behavior from an untrained brain-and-nerve-cord connectome
- 2610.40344-Morphological Features of Murine Pulmonary Arterial Networks for Diagnosing Vascular Disease
- Daily Summary for 2026-11-19
- 2609.10193-Are You Learning Biological Signal or Shortcuts? Auditing and Mitigating Bias in Protein-Protein Interaction Datasets
- 2610.38236-Generative discovery of chemical reaction networks
- 2507.11176-An Interpretable AI framework Quantifying Traditional Chinese Medicine Principles Towards Enhancing and Integrating with Modern Biomedicine
- The day in conversation for 2026-11-19
- Daily Summary for 2026-11-19
- 2504.03590-Towards a Unified Framework for Determining Conformational Ensembles of Disordered Proteins
- 2603.15217-A multiscale discrete-to-continuum framework for structured population models