Inference of population genetics parameters using discriminator neural networks: an adversarial Monte Carlo approach

Gower, Graham; Picazo, Pablo Iáñez; Lindgren, Finn; Racimo, Fernando

doi:10.1101/2023.04.27.538386

Cited by 5 publications

(2 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nevertheless, our experiments with pg-gan were conducted without specialized neural network hardware and do not dismiss GANs' potential as an emerging approach. Further training and improved procedures may enhance GAN-based demographic inference 22 .…”

Section: Methods Overview: Genealogical Likelihood Under Multi-popula...mentioning

confidence: 99%

A likelihood-based framework for demographic inference from genealogical trees

Fan,

Cahoon,

Dinh

et al. 2023

Preprint

View full text Add to dashboard Cite

The demographic history of a population drives the pattern of genetic variation and is encoded in the gene-genealogical trees of the sampled alleles. However, existing methods to infer demographic history from genetic data tend to use relatively low-dimensional summaries of the genealogy, such as allele frequency spectra. As a step toward capturing more of the information encoded in the genome-wide sequence of genealogical trees, here we propose a novel framework called the genealogical likelihood (gLike), which derives the full likelihood of a genealogical tree under any hypothesized demographic history. Employing a graph-based structure, gLike summarizes across independent trees the relationships among all lineages in a tree with all possible trajectories of population memberships through time and efficiently computes the exact marginal probability under a parameterized demographic model. Through extensive simulations and empirical applications on populations that have experienced multiple admixtures, we showed that gLike can accurately estimate dozens of demographic parameters when the true genealogy is known, including ancestral population sizes, admixture timing, and admixture proportions. Moreover, when using genealogical trees inferred from genetic data, we showed that gLike outperformed conventional demographic inference methods that leverage only the allele-frequency spectrum and yielded parameter estimates that align with established historical knowledge of the past demographic histories for populations like Latino Americans and Native Hawaiians. Furthermore, our framework can trace ancestral histories by analyzing a sample from the admixed population without proxies for its source populations, removing the need to sample ancestral populations that may no longer exist. Taken together, our proposed gLike framework harnesses underutilized genealogical information to offer exceptional sensitivity and accuracy in inferring complex demographies for humans and other species, particularly as estimation of genome-wide genealogies improves.

show abstract

Section: Methods Overview: Genealogical Likelihood Under Multi-popula...mentioning

confidence: 99%

A likelihood-based framework for demographic inference from genealogical trees

Fan,

Cahoon,

Dinh

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The key innovation of is that it learns an explicit evolutionary generative model, in contrast to other GANs which generate sequences that look like real data from random processes with no underlying model ( Yelmen et al 2021 ; Booker et al 2023 ). Recent work has made use of for other species ( Small et al 2023 ) and improved the approach using adversarial Monte Carlo methods ( Gower et al 2023 ).…”

Section: Introductionmentioning

confidence: 99%

Interpreting generative adversarial networks to infer natural selection from genetic data

Riley,

Mathieson,

Mathieson

2024

Genetics

View full text Add to dashboard Cite

Understanding natural selection and other forms of non-neutrality is a major focus for the use of machine learning in population genetics. Existing methods rely on computationally intensive simulated training data. Unlike efficient neutral coalescent simulations for demographic inference, realistic simulations of selection typically require slow forward simulations. Because there are many possible modes of selection, a high dimensional parameter space must be explored, with no guarantee that the simulated models are close to the real processes. Finally, it is difficult to interpret trained neural networks, leading to a lack of understanding about what features contribute to classification. Here we develop a new approach to detect selection and other local evolutionary processes that requires relatively few selection simulations during training. We build upon a Generative Adversarial Network (GAN) trained to simulate realistic neutral data. The resulting GAN consists of a generator (fitted demographic model), and a discriminator (convolutional neural network) that predicts whether a genomic region is real or fake. As the generator can only generate data under neutral demographic processes, regions of real data that the discriminator recognizes as having a high probability of being "real" do not fit the neutral demographic model and are therefore candidates for targets of selection. To incentivize identification of a specific mode of selection, we fine-tune the discriminator with a small number of custom non-neutral simulations. We show that this approach has high power to detect various forms of selection in simulations, and that it finds regions under positive selection identified by state-of-the-art population genetic methods in three human populations. Finally, we show how to interpret the trained networks by clustering hidden units of the discriminator based on their correlation patterns with known summary statistics.

show abstract