[Paper Review] Human Ancestries Simulation and Inference: a Review of Ancestral Recombination Graph Samplers

Patrick Fournier, Fabrice Larribe|arXiv (Cornell University)|Jan 14, 2026

Genetic Associations and Epidemiology0 citations

TL;DR

A comprehensive technical review of 32 Arg samplers (simulation and inference) from early 2000s to present, focusing on performance, usability, and biological realism to guide researchers in implementing coalescent-with-recombination samplers.

ABSTRACT

There is little debate about the importance of the ancestral recombination graph in population genetics. An important theoretical tool, the main obstacle to its widespread usage is the computational cost required to match the ever-increasing scale of the data being analyzed. Many of these difficulties have been overcome in the past two decades, which have consequently seen the development of increasingly sophisticated ARG simulation and inference software. Nonetheless, challenges remain, especially in the area of ancestry inference. This paper is a comprehensive review of ARG samplers that have emerged in the past three decades to meet the need for scalable and flexible ancestry simulation and inference solutions. It specifically focuses on their performance, usability, and the biological realism of the underlying algorithm, and aims primarily to provide a technical overview of the field for researchers seeking to write their own coalescent-with-recombination sampler. As a complement to this article, we have compiled links to software, source code and documentation and made them available at https://www.patrickfournier.ca/arg-samplers-review/graph.

Motivation & Objective

• Survey and compare ARG samplers developed over three decades (2000s–present).
• Evaluate performance, usability, and biological realism of each sampler.
• Differentiate model-based vs heuristic approaches and their applicability to simulation and inference.
• Provide guidance for researchers aiming to implement their own coalescent-with-recombination sampler.

Proposed method

• Classify samplers into model-based versus heuristic-based families and distinguish simulated versus inferred outputs.
• Analyze event sampling strategies and their impact on accuracy and performance, including coalescence and recombination event types.
• Discuss approximation schemes (e.g., SMC/SMC') and their statistical implications.
• Summarize programming languages, interfaces, and practical considerations for software implementation.
• Present a detailed, feature-focused comparison of 32 samplers and related literature.

Experimental results

Research questions

• RQ1 What are the main design philosophies (model-based vs heuristic) used by ARG samplers, and how do they affect accuracy and efficiency?
• RQ2 How do different samplers handle coalescence and recombination event types, and what are the trade-offs between realism and performance?
• RQ3 What are the practical performance implications of using exact Hudson-based algorithms versus approximate SMC/SMC' approaches?
• RQ4 How do implementation choices (language, interface) influence usability and extensibility for researchers building or modifying samplers?
• RQ5 What criteria should guide the selection or development of ARG samplers for simulation vs inference tasks?

Key findings

• Hudson’s original algorithm (ms) remains a reference implementation for exact CWR sampling, with broad feature support but limited genome-scale performance.
• msprime introduces a Tree Sequence data structure enabling genome-scale performance improvements over ms by exploiting marginal tree correlations, allowing simulations without full ARG storage.
• Approximate spatial samplers (SMC/SMC') trade realism for linear-time performance by constraining recombination event locations, with varying accuracy depending on demographic structure.
• Recombination event types and their relation to ancestral material influence sampler design; some approaches trade off sampling type 2 events for computational efficiency.
• Simulation flexibility has been extended to include selection, migration, gene conversion, and ancient DNA, with performance driven by event reduction strategies and data structures.
• The review provides extensive benchmarking and practical guidance to researchers aiming to implement or modify coalescent-with-recombination samplers.