Boosting forward-time population genetic simulators through genotype compression

Abstract

BackgroundForward-time population genetic simulations play a central role in deriving and testing evolutionary hypotheses. Such simulations may be data-intensive, depending on the settings to the various parameters controlling them. In particular, for certain settings, the data footprint may quickly exceed the memory of a single compute node.ResultsWe develop a novel and general method for addressing the memory issue inherent in forward-time simulations by compressing and decompressing, in real-time, active and ancestral genotypes, while carefully accounting for the time overhead. We propose a general graph data structure for compressing the genotype space explored during a simulation run, along with efficient algorithms for constructing and updating compressed genotypes which support both mutation and recombination. We tested the performance of our method in very large-scale simulations. Results show that our method not only scales well, but that it also overcomes memory issues that would cripple existing tools.ConclusionsAs evolutionary analyses are being increasingly performed on genomes, pathways, and networks, particularly in the era of systems biology, scaling population genetic simulators to handle large-scale simulations is crucial. We believe our method offers a significant step in that direction. Further, the techniques we provide are generic and can be integrated with existing population genetic simulators to boost their performance in terms of memory usage.

A spatially aware likelihood test to detect sweeps from haplotype distributions

“…All neutral and selection simulations were run for 11 N generations, where the first 10 N generations were used as burn-in and n = 50 diploid individuals were sampled from the population after 11 N generations ( i.e ., the present). Because forward-time simulations are computationally intensive, as is commonly-practiced [ 89 , 90 ] we scaled all constant-size demographic history simulations by a factor λ = 10 and the European human history by λ = 20, such that the selection coefficient, mutation rate, and recombination rate were multiplied by λ and the population size at each generation and the total number of simulated generations were divided by λ. This scaling leads to a speedup of approximately λ 2 in computing time, such that the constant-size simulations run roughly 100 times faster than without scaling and the CEU model simulations run approximately 400 times faster, making a large-scale simulation study feasible.…”

“…We also examined the application of Λ, T , and G123 (analogue of H12 [ 10 ]) to unphased multilocus genotype input data to evaluate the relative powers of these three approaches when applied on study systems for which obtaining phased haplotypes is difficult, unreliable, or impossible [ 91 ]. We applied the software released with this article for application of the Λ statistic, the T statistic, and H12 (and G123), and the software [ 90 ] to compute standardized iHS and nS L .…”

An overview of the utility of population simulation software in molecular ecology

“…Such models can link to social and economic (Haight et al 2002) or public health models (Magori et al 2009), can be applied to managing genetic resources of domesticated species and may inform emerging issues like assisted colonization. Recent advances in data compression (Ruths & Nakhleh 2013) and data reduction (Aberer & Stamatakis 2013) hold particular promise for forward simulation of large sequences (>10 6 base pairs) for large numbers of individuals (N > 10 5 ), which are expensive for time and memory. The modular nature of new simulators also makes them easily expandable with new features Rebaudo et al 2013;Strand & Niehaus 2007).…”