IBD & ROH: What They Reveal About Your Cohort
This guide provides a defensible, decision-ready workflow for IBD/ROH interpretation and integrates smoothly with our Genetic Diversity service and Population Structure Analysis workflows.
TL;DR — Read IBD and ROH Like a Decision Tree
If long ROH (e.g., >5–10 Mb) dominate an individual's genome, you're looking at recent inbreeding; quantify F_ROH and update mating plans. A burden of short ROH (≈0.5–1 Mb) signals older drift or a historically small effective population size (Ne); track trajectories, not just snapshots. Widespread IBD segments between pairs point to recent shared ancestry; feed those segments to IBDNe to estimate recent Ne (≈5–50 generations) and set recovery targets. Recurrent ROH islands across individuals mark putative selection or recombination deserts—interrogate before acting. Together, IBD and ROH convert pedigrees and genotypes into actionable levers for genetic management in herds, breeds, and conservation programs.
Why This Matters Now: The Hidden Cost of Missing Realized Inbreeding
Pedigrees summarize expected relatedness; the genome records what actually happened. Runs of homozygosity (ROH) arise when both chromosomal copies descend from a shared ancestor, producing long homozygous tracts. The number and length of ROH reveal recent versus ancient common ancestry, while the genome-wide proportion F_ROH captures realized inbreeding that pedigrees often miss—especially in incomplete or shallow pedigree systems common in livestock and wildlife. In practice, F_ROH correlates more closely with fitness outcomes than pedigree-based F, making it the more reliable early-warning signal for inbreeding depression and suboptimal crosses.
For conservation programs balancing demographic recovery with local adaptation, the stakes are high: management that ignores realized inbreeding risks lower survival, fertility, and resilience. Genomic metrics enable you to detect, quantify, and date inbreeding, so breeding plans and translocations can be tuned with evidence rather than guesswork.
What IBD and ROH Deliver (Beyond Pedigrees)
Quantify Realized Inbreeding with F_ROH
F_ROH is the fraction of the autosome in ROH per individual. When tracked alongside fertility, growth, or health outcomes, it clarifies which animals carry the highest realized inbreeding and which matings are most prudent. F_ROH is routinely more informative than pedigree-based estimates in herds with missing or shallow records.
Separate "Recent" Versus "Historical" Inbreeding
ROH length classes are time-aware indicators: long ROH reflect recent common ancestors (a few generations), whereas short ROH reflect older drift from a historically small Ne. This distinction informs whether managers should adjust near-term matings or expand founders/immigrants to lift long-term diversity.
Reconstruct Recent Effective Population Size From IBD
Identity-by-descent (IBD) segments shared between individuals carry information about recent Ne. With IBDNe, you can infer Ne(t) over roughly the last 5–50 generations, quantify whether a population is declining, stable, or recovering, and benchmark the effect of management interventions.
Flag ROH Islands and Candidate Selection Signals
ROH islands—regions repeatedly homozygous across many individuals—can mark selection, drift, or recombination deserts. Cross-check islands against recombination maps and trait associations before acting; livestock syntheses show islands in known targets such as MSTN and other production-related loci. Treat these as candidates for validation, not definitive proof of selection.
Methods That Stand Up to Review (and Scale)
ROH calling on sequence and arrays. For whole-genome sequencing or exomes, BCFtools/RoH uses a hidden Markov model (HMM) to detect autozygous tracts, handling genotype likelihoods gracefully. For arrays or rapid screening, PLINK provides fast sliding-window ROH detection with transparent parameters (--homozyg-*). To partition autozygosity by age classes, RZooRoH models HBD (homozygosity-by-descent) classes, profiling recent and ancient contributions within individuals.
IBD detection for pairs. Refined IBD shows improved power and accuracy for phased data and is a robust source of segment calls for IBDNe. Together they deliver interpretable recent effective population size trajectories.
Ancient or low-coverage samples. for archaeological or sparse modern data, hapROH leverages reference haplotypes to call long ROH at very low coverage, enabling parental-relatedness inference through time.
Detecting runs of homozygosity using a reference panel. (Ringbauer H. et al. (2021) Nature Communications).
Step-By-Step Workflow: From VCF to Decisions
1) Data Preparation and QC
- Harmonize reference build, strand, and variant IDs across cohorts.
- Filter variants for call rate and Hardy–Weinberg expectations appropriate to study design.
- Verify sex checks and contamination flags before ROH/IBD to avoid false signals.
- Choose your ROH tool by data type: BCFtools/RoH for sequence/exome; PLINK for arrays and rapid cohort scans; RZooRoH when you need age-class decomposition for nuanced interpretation.
This stage feeds directly into our Genetic Diversity service, where QC thresholds and parameter templates are standardized so recurring projects remain comparable across seasons and sites.
2) Call ROH and Compute F_ROH
- Set minimum ROH length thresholds mindful of species and marker density (e.g., start at 0.5 Mb for sequence, adjust upward on sparse arrays).
- Bin ROH by length (e.g., 0.5–1 Mb, 1–5 Mb, >5–10 Mb) to create a within-individual time profile.
- Compute F_ROH as Σ(ROH length) / autosomal length per individual; summarize by group (line, subpopulation, release cohort) to identify at-risk clusters and candidates for outcrossing.
When F_ROH is high and dominated by long ROH, prioritize near-term mating changes. When F_ROH is high but dominated by short ROH, treat it as a legacy of historically small Ne and consider founder expansion or genetic rescue rather than just reshuffling current pairings.
Total length and number of ROH for three size classes including Small (0.5 to 1 Mb), Medium (1 to 5 Mb) and Large (>5 Mb). (Liu C. et al. (2022) PLOS ONE).
3) Detect IBD Segments Between Individuals
- Phase genotypes (if needed), then run Refined IBD on autosomes.
- Filter segments below a conservative size (e.g., ≥2–3 cM) to limit false positives; document the genetic map used.
- Summarize per-pair IBD counts/lengths and visualize network structure (e.g., kinship graphs) to highlight unexpected close relationships that pedigrees missed.
Relatedness that slips through QC inflates ROH in offspring and masks true population structure—a problem for both breeding value estimation and conservation triage.
4) Estimate Recent Ne with IBDNe
- Input: filtered IBD segments (optionally ancestry-specific).
- Run: IBDNe to obtain Ne(t) trajectories with confidence intervals.
- Diagnose: categorize trends (decline, stability, recovery) and set numeric targets that align with your management horizon (e.g., "double Ne within 10 generations").
For multi-origin herds or reintroduced populations, estimate ancestry-specific Ne to catch uneven recovery that a pooled analysis could mask.
5) Identify ROH Islands and Candidate Selection
- Aggregate ROH overlap across individuals to find ROH islands; annotate with gene content and recombination rate.
- Interpret cautiously: islands may reflect selection or reduced recombination; compare across related populations and triangulate with trait/fitness data or GWAS to avoid false leads.
- Prioritize validation: if islands intersect known candidate genes or QTLs, plan targeted follow-up (e.g., haplotype diversity checks, functional assays).
Incidence plots of SNPs in ROH for six sheep populations. (Gorssen W. et al. (2021) Genetics Selection Evolution).
6) Decide and Document
- Near-term mating plan: use F_ROH and pairwise IBD together to avoid close crosses, manage line contributions, and preserve rare haplotypes.
- Population targets: use IBDNe to motivate founder additions or immigration, with monitoring every 1–2 breeding cycles.
- Reporting: include parameter blocks (software, versions, thresholds), figures (ROH spectra, Ne plots), and tables (per-sample F_ROH, pairwise IBD) in the reviewer-ready pack for boards, funders, and journals.
These deliverables are included in our genetic diversity analysis so you can re-run the same analysis year-over-year and demonstrate improvement quantitatively.
Interpreting Patterns: Concrete Scenarios and Actions
Long ROH Are Common and F_ROH Is High
Meaning: recent inbreeding (close matings or small current Ne).
Action: update mating design, broaden crosses, and monitor F_ROH in the next generation. In conservation contexts, plan genetic rescue to inject diversity.
Short ROH Are Abundant, But Long ROH Are Rare
Meaning: legacy of historically small Ne with little current inbreeding.
Action: protect effective size—avoid severe skews in breeder contributions, maintain census size, and consider immigration over within-line reshuffling.
Pairwise IBD Is Widespread; IBDNe Suggests Declining Ne
Meaning: population contraction over the last 5–50 generations.
Action: set numeric Ne targets, plan founder recruitment, and re-estimate Ne every cycle to validate improvement.
Plots of GAM models, with the variable to be predicted "AV index" and the predictors. (Blondeau Da Silva S. et al. (2024) Heredity).
ROH Islands Align with Known Traits
Meaning: probable selection (e.g., production traits in livestock).
Action: moderate selection intensity or leverage alternative lines to retain diversity around targeted loci; cross-check with recombination maps to avoid mistaking recombination deserts for selection.
Scatter plot of total homozygosity per individual against four corrected carcass traits. (Zhao X. et al. (2021) BMC Genomics)
Low Coverage or Historical Samples
Meaning: standard callers may fail.
Action: use hapROH to recover long ROH; interpret results within a time-aware framework (e.g., by archaeological period or release cohorts).
Reporting Templates
Methods (example wording):
"We performed ROH calling with BCFtools/RoH (sequence) and PLINK (arrays) using species-appropriate thresholds; for age-class partitioning we applied RZooRoH. We detected pairwise IBD in phased data with Refined IBD, and inferred recent effective population size using IBDNe from filtered segments. We computed F_ROH per individual, summarized ROH length bins, estimated Ne(t) with confidence intervals, and identified ROH islands by overlapping ROH across individuals. Software versions, recombination maps, and parameter sets are documented in Supplementary Methods."
Figures:
- ROH length spectra by group (short/medium/long).
- F_ROH distribution (violin or ridge) with management thresholds.
- Pairwise IBD network highlighting unexpected kin.
- IBDNe trajectory with annotations (management events, releases).
- ROH island Manhattan plot with candidate genes annotated.
Tables: per-sample F_ROH, pairwise IBD summaries, IBDNe outputs, island coordinates and gene overlaps.
Quality Checks and Pitfalls to Avoid
- Marker Density Bias: sparse arrays under-call short ROH; sequence-based HMM callers reduce error. Report platform limits and perform sensitivity analyses on the minimum ROH length.
- Threshold Portability: ROH cutoffs are species- and map-dependent; avoid copy-pasting human thresholds to cattle, dogs, or endangered wildlife without calibration. Use RZooRoH to explore age-class partitions when uncertain.
- IBD False Positives: mis-phasing or genotyping errors create spurious short segments; use conservative cM cutoffs and quality filters. Refined IBD generally shows favorable power/accuracy.
- Over-Interpreting ROH Islands: recurrent ROH can reflect selection or recombination deserts; triangulate with recombination rate, cross-population comparison, and trait/fitness associations before acting.
- Narrative Overreach: if PCA or fineSTRUCTURE disagrees with ROH/IBD, revisit batch effects, map build, and phasing rather than selectively reporting concordant results.
FAQ — Practical, Answer-First Guidance
Start with published ranges (e.g., 0.5–1 Mb = older drift, 1–5 Mb = intermediate, >5–10 Mb = recent) and adjust for marker density and species-specific recombination. Keep bins consistent over time and explain them in Methods.
Yes. Across simulations and real populations, marker-based measures capture realized IBD more accurately, especially when pedigrees are shallow or incomplete. For management, F_ROH is typically the more predictive lever.
IBDNe infers Ne over about 5–50 generations from the distribution of long IBD segments. It is well suited to the timeframe of modern breeding and conservation actions.
Use BCFtools/RoH (HMM) for sequence/exome; PLINK for quick array-based scans; RZooRoH to decompose HBD into age classes; Refined IBD + IBDNe to add recent Ne context. Document versions, maps, and thresholds for reproducibility.
Yes. hapROH reliably detects long ROH from low-coverage data using a modern reference panel and has been applied to thousands of ancient genomes to infer parental relatedness.
Conclusion and Next Steps
IBD and ROH are practical genomic instruments for protecting population health. By combining F_ROH, ROH length spectra, pairwise IBD, IBDNe, and ROH islands, you can separate recent inbreeding from legacy drift, quantify recent Ne, and target interventions where they matter. The playbook is reproducible and audit-ready: choose the right caller (BCFtools/RoH, PLINK, RZooRoH), detect and filter IBD with Refined IBD, infer Ne(t) with IBDNe, and ship a clear Methods/reporting pack.
If you want these insights operationalized with consistent deliverables, our genetic diversity analysis packages ROH/IBD calling → F_ROH & IBDNe → reviewer-ready reporting, and integrates with population structure analysis for stratification control. We'll calibrate thresholds by species and platform, establish ROH length bins, set Ne targets, and hand you a repeatable SOP you can run every season.
Related reading:
References
- Ringbauer, H., Novembre, J., Steinrücken, M. et al. Parental relatedness through time revealed by runs of homozygosity in ancient DNA. Nature Communications 12, 5425 (2021).
- Gorssen, W., Meyermans, R., Janssens, S. et al. A publicly available repository of ROH islands reveals signatures of selection in different livestock and pet species. Genetics Selection Evolution 53, 2 (2021).
- Liu, Y., Zhao, G., Lin, X. et al. Genomic inbreeding and runs of homozygosity analysis of indigenous cattle populations in southern China. PLOS ONE 17, e0271718 (2022).
- Blondeau Da Silva, S., Mwacharo, J.M., Li, M. et al. IBD sharing patterns as intra-breed admixture indicators in small ruminants. Heredity 132, 30–42 (2024).
- Zhao, G., Liu, Y., Niu, Q. et al. Runs of homozygosity analysis reveals consensus homozygous regions affecting production traits in Chinese Simmental beef cattle. BMC Genomics 22, 678 (2021).
- Browning, S.R., Browning, B.L. Accurate non-parametric estimation of recent effective population size from segments of identity by descent. The American Journal of Human Genetics 97, 404–418 (2015).
- Narasimhan, V., Danecek, P., Scally, A. et al. BCFtools/RoH: a hidden Markov model approach for detecting autozygosity from next-generation sequencing data. Bioinformatics 32, 1749–1751 (2016).
Related
Services
Download
