Efficient pedigree recording for fast population genetics simulation

• DOI: 10.1371/journal.pcbi.1006581
• Publication Date: 2018-11-01T17:29:47Z
• Authors: Jerome Kelleher; Kevin R. Thornton; Jaime Ashander; Peter L. Ralph
• Keywords: Population Genetics; Human evolutionary genetics

In this paper we describe how to efficiently record the entire genetic history of a population in forwards-time, individual-based genetics simulations with arbitrary breeding models, structure and demography. This approach dramatically reduces computational burden tracking individual genomes by allowing us simulate only those loci that may affect reproduction (those having non-neutral variants). The is recorded as succinct tree sequence introduced software package msprime, on which neutral mutations can be quickly placed afterwards. Recording results each event requires storage grows linearly time, but there great deal redundancy information. We solve problem providing an algorithm 'simplify' removing irrelevant for given set genomes. By periodically simplifying respect extant population, show total space required modest overall large efficiency gains made over classical forward-time simulations. implement general-purpose framework recording genealogical data, used make any model more efficient. modify two popular forwards-time simulation frameworks use new observe large, whole-genome one orders magnitude. addition speed, our method pedigrees has several advantages: (1) All marginal genealogies simulated individuals are recorded, rather than just genotypes. (2) A N M polymorphic sites stored O(N log + M) space, making it feasible store simulation's final generation well its history. (3) easily initialized efficient coalescent deep processing sequences named tskit.