Supervised Machine Learning for Population Genetics: A New Paradigm

Artificial intelligence enables unified analysis of historical and landscape influences on genetic diversity

While genetic variation in any species is potentially shaped by a range of processes, phylogeography and landscape genetics are largely concerned with inferring how environmental conditions and landscape features impact neutral intraspecific diversity. However, even as both disciplines have come to utilize SNP data over the last decades, analytical approaches have remained for the most part focused on either broad-scale inferences of historical processes (phylogeography) or on more localized inferences about environmental and/or landscape features (landscape genetics). Here we demonstrate that an artificial intelligence model-based analytical framework can consider both deeper historical factors and landscape-level processes in an integrated analysis. We implement this framework using data collected from two Brazilian anurans, the Brazilian sibilator frog (Leptodactylus troglodytes) and granular toad (Rhinella granulosa). Our results indicate that historical demographic processes shape most the genetic variation in the sibulator frog, while landscape processes primarily influence variation in the granular toad. The machine learning framework used here allows both historical and landscape processes to be considered equally, rather than requiring researchers to make an a priori decision about which factors are important.

Beyond the hype: using AI, big data, wearable devices, and the internet of things for high-throughput livestock phenotyping

Phenotyping of animals is a routine task in agriculture which can provide large datasets for the functional annotation of genomes. Using the livestock farming sector to study complex traits enables genetics researchers to fully benefit from the digital transformation of society as economies of scale substantially reduces the cost of phenotyping animals on farms. In the agricultural sector genomics has transitioned towards a model of 'Genomics without the genes' as a large proportion of the genetic variation in animals can be modelled using the infinitesimal model for genomic breeding valuations. Combined with third generation sequencing creating pan-genomes for livestock the digital infrastructure for trait collection and precision farming provides a unique opportunity for high-throughput phenotyping and the study of complex traits in a controlled environment. The emphasis on cost efficient data collection mean that mobile phones and computers have become ubiquitous for cost-efficient large-scale data collection but that the majority of the recorded traits can still be recorded manually with limited training or tools. This is especially valuable in low- and middle income countries and in settings where indigenous breeds are kept at farms preserving more traditional farming methods. Digitalization is therefore an important enabler for high-throughput phenotyping for smaller livestock herds with limited technology investments as well as large-scale commercial operations. It is demanding and challenging for individual researchers to keep up with the opportunities created by the rapid advances in digitalization for livestock farming and how it can be used by researchers with or without a specialization in livestock. This review provides an overview of the current status of key enabling technologies for precision livestock farming applicable for the functional annotation of genomes.

Tree sequences as a general-purpose tool for population genetic inference

As population genetics data increases in size new methods have been developed to store genetic information in efficient ways, such as tree sequences. These data structures are computationally and storage efficient, but are not interchangeable with existing data structures used for many population genetic inference methodologies such as the use of convolutional neural networks (CNNs) applied to population genetic alignments. To better utilize these new data structures we propose and implement a graph convolutional network (GCN) to directly learn from tree sequence topology and node data, allowing for the use of neural network applications without an intermediate step of converting tree sequences to population genetic alignment format. We then compare our approach to standard CNN approaches on a set of previously defined benchmarking tasks including recombination rate estimation, positive selection detection, introgression detection, and demographic model parameter inference. We show that tree sequences can be directly learned from using a GCN approach and can be used to perform well on these common population genetics inference tasks with accuracies roughly matching or even exceeding that of a CNN-based method. As tree sequences become more widely used in population genetics research we foresee developments and optimizations of this work to provide a foundation for population genetics inference moving forward.

Artificial intelligence in plant breeding

Harnessing cutting-edge technologies to enhance crop productivity is a pivotal goal in modern plant breeding. Artificial intelligence (AI) is renowned for its prowess in big data analysis and pattern recognition, and is revolutionizing numerous scientific domains including plant breeding. We explore the wider potential of AI tools in various facets of breeding, including data collection, unlocking genetic diversity within genebanks, and bridging the genotype-phenotype gap to facilitate crop breeding. This will enable the development of crop cultivars tailored to the projected future environments. Moreover, AI tools also hold promise for refining crop traits by improving the precision of gene-editing systems and predicting the potential effects of gene variants on plant phenotypes. Leveraging AI-enabled precision breeding can augment the efficiency of breeding programs and holds promise for optimizing cropping systems at the grassroots level. This entails identifying optimal inter-cropping and crop-rotation models to enhance agricultural sustainability and productivity in the field.

Selection leads to false inferences of introgression using popular methods

Detecting introgression between closely related populations or species is a fundamental objective in evolutionary biology. Existing methods for detecting migration and inferring migration rates from population genetic data often assume a neutral model of evolution. Growing evidence of the pervasive impact of selection on large portions of the genome across diverse taxa suggests that this assumption is unrealistic in most empirical systems. Further, ignoring selection has previously been shown to negatively impact demographic inferences (e.g. of population size histories). However, the impacts of biologically realistic selection on inferences of migration remain poorly explored. Here, we simulate data under models of background selection, selective sweeps, balancing selection, and adaptive introgression. We show that ignoring selection sometimes leads to false inferences of migration in popularly used methods that rely on the site frequency spectrum. Specifically, balancing selection and some models of background selection result in the rejection of isolation-only models in favor of isolation-with-migration models and lead to elevated estimates of migration rates. BPP, a method that analyzes sequence data directly, showed false positives for all conditions at recent divergence times, but balancing selection also led to false positives at medium-divergence times. Our results suggest that such methods may be unreliable in some empirical systems, such that new methods that are robust to selection need to be developed.

Fast and Accurate Estimation of Selection Coefficients and Allele Histories from Ancient and Modern DNA

We here present CLUES2, a full-likelihood method to infer natural selection from sequence data that is an extension of the method CLUES. We make several substantial improvements to the CLUES method that greatly increases both its applicability and its speed. We add the ability to use ancestral recombination graphs on ancient data as emissions to the underlying hidden Markov model, which enables CLUES2 to use both temporal and linkage information to make estimates of selection coefficients. We also fully implement the ability to estimate distinct selection coefficients in different epochs, which allows for the analysis of changes in selective pressures through time, as well as selection with dominance. In addition, we greatly increase the computational efficiency of CLUES2 over CLUES using several approximations to the forward-backward algorithms and develop a new way to reconstruct historic allele frequencies by integrating over the uncertainty in the estimation of the selection coefficients. We illustrate the accuracy of CLUES2 through extensive simulations and validate the importance sampling framework for integrating over the uncertainty in the inference of gene trees. We also show that CLUES2 is well-calibrated by showing that under the null hypothesis, the distribution of log-likelihood ratios follows a χ2 distribution with the appropriate degrees of freedom. We run CLUES2 on a set of recently published ancient human data from Western Eurasia and test for evidence of changing selection coefficients through time. We find significant evidence of changing selective pressures in several genes correlated with the introduction of agriculture to Europe and the ensuing dietary and demographic shifts of that time. In particular, our analysis supports previous hypotheses of strong selection on lactase persistence during periods of ancient famines and attenuated selection in more modern periods.

High dispersal ability versus migratory traditions: Fine-scale population structure and post-glacial colonisation in bar-tailed godwits

In migratory animals, high mobility may reduce population structure through increased dispersal and enable adaptive responses to environmental change, whereas rigid migratory routines predict low dispersal, increased structure, and limited flexibility to respond to change. We explore the global population structure and phylogeographic history of the bar-tailed godwit, Limosa lapponica, a migratory shorebird known for making the longest non-stop flights of any landbird. Using nextRAD sequencing of 14,318 single-nucleotide polymorphisms and scenario-testing in an Approximate Bayesian Computation framework, we infer that bar-tailed godwits existed in two main lineages at the last glacial maximum, when much of their present-day breeding range persisted in a vast, unglaciated Siberian-Beringian refugium, followed by admixture of these lineages in the eastern Palearctic. Subsequently, population structure developed at both longitudinal extremes: in the east, a genetic cline exists across latitude in the Alaska breeding range of subspecies L. l. baueri; in the west, one lineage diversified into three extant subspecies L. l. lapponica, taymyrensis, and yamalensis, the former two of which migrate through previously glaciated western Europe. In the global range of this long-distance migrant, we found evidence of both (1) fidelity to rigid behavioural routines promoting fine-scale geographic population structure (in the east) and (2) flexibility to colonise recently available migratory flyways and non-breeding areas (in the west). Our results suggest that cultural traditions in highly mobile vertebrates can override the expected effects of high dispersal ability on population structure, and provide insights for the evolution and flexibility of some of the world's longest migrations.

Modelling the demographic history of human North African genomes points to a recent soft split divergence between populations

BACKGROUND

North African human populations present a complex demographic scenario due to the presence of an autochthonous genetic component and population substructure, plus extensive gene flow from the Middle East, Europe, and sub-Saharan Africa.

RESULTS

We conducted a comprehensive analysis of 364 genomes to construct detailed demographic models for the North African region, encompassing its two primary ethnic groups, the Arab and Amazigh populations. This was achieved through an Approximate Bayesian Computation with Deep Learning (ABC-DL) framework and a novel algorithm called Genetic Programming for Population Genetics (GP4PG). This innovative approach enabled us to effectively model intricate demographic scenarios, utilizing a subset of 16 whole genomes at > 30X coverage. The demographic model suggested by GP4PG exhibited a closer alignment with the observed data compared to the ABC-DL model. Both point to a back-to-Africa origin of North African individuals and a close relationship with Eurasian populations. Results support different origins for Amazigh and Arab populations, with Amazigh populations originating back in Epipaleolithic times, while GP4PG supports Arabization as the main source of Middle Eastern ancestry. The GP4PG model includes population substructure in surrounding populations (sub-Saharan Africa and Middle East) with continuous decaying gene flow after population split. Contrary to ABC-DL, the best GP4PG model does not require pulses of admixture from surrounding populations into North Africa pointing to soft splits as drivers of divergence in North Africa.

CONCLUSIONS

We have built a demographic model on North Africa that points to a back-to-Africa expansion and a differential origin between Arab and Amazigh populations.

Discovering Fragile Clades and Causal Sequences in Phylogenomics by Evolutionary Sparse Learning

Phylogenomic analyses of long sequences, consisting of many genes and genomic segments, reconstruct organismal relationships with high statistical confidence. But, inferred relationships can be sensitive to excluding just a few sequences. Currently, there is no direct way to identify fragile relationships and the associated individual gene sequences in species. Here, we introduce novel metrics for gene-species sequence concordance and clade probability derived from evolutionary sparse learning models. We validated these metrics using fungi, plant, and animal phylogenomic datasets, highlighting the ability of the new metrics to pinpoint fragile clades and the sequences responsible. The new approach does not necessitate the investigation of alternative phylogenetic hypotheses, substitution models, or repeated data subset analyses. Our methodology offers a streamlined approach to evaluating major inferred clades and identifying sequences that may distort reconstructed phylogenies using large datasets.

Applications of machine learning in phylogenetics

Machine learning has increasingly been applied to a wide range of questions in phylogenetic inference. Supervised machine learning approaches that rely on simulated training data have been used to infer tree topologies and branch lengths, to select substitution models, and to perform downstream inferences of introgression and diversification. Here, we review how researchers have used several promising machine learning approaches to make phylogenetic inferences. Despite the promise of these methods, several barriers prevent supervised machine learning from reaching its full potential in phylogenetics. We discuss these barriers and potential paths forward. In the future, we expect that the application of careful network designs and data encodings will allow supervised machine learning to accommodate the complex processes that continue to confound traditional phylogenetic methods.

Integration of machine learning models with microsatellite markers: New avenue in world grapevine germplasm characterization

Development of efficient analytical techniques is required for effective interpretation of biological data to take novel hypotheses and finding the critical predictive patterns. Machine Learning algorithms provide a novel opportunity for development of low-cost and practical solutions in biology. In this study, we proposed a new integrated analytical approach using supervised machine learning algorithms and microsatellites data of worldwide vitis populations. A total of 1378 wild (V. vinifera spp. sylvestris) and cultivated (V. vinifera spp. sativa) accessions of grapevine were investigated using 20 microsatellite markers. Data cleaning, feature selection, and supervised machine learning classification models vis, Naive Bayes, Support Vector Machine (SVM) and Tree Induction methods were implied to find most indicative and diagnostic alleles to represent wild/cultivated and originated geography of each population. Our combined approaches showed microsatellite markers with the highest differentiating capacity and proved efficiency for our pipeline of classification and prediction of vitis accessions. Moreover, our study proposed the best combination of markers for better distinguishing of populations, which can be exploited in future germplasm conservation and breeding programs.

Phase-type distributions in mathematical population genetics: An emerging framework

A phase-type distribution is the time to absorption in a continuous- or discrete-time Markov chain. Phase-type distributions can be used as a general framework to calculate key properties of the standard coalescent model and many of its extensions. Here, the 'phases' in the phase-type distribution correspond to states in the ancestral process. For example, the time to the most recent common ancestor and the total branch length are phase-type distributed. Furthermore, the site frequency spectrum follows a multivariate discrete phase-type distribution and the joint distribution of total branch lengths in the two-locus coalescent-with-recombination model is multivariate phase-type distributed. In general, phase-type distributions provide a powerful mathematical framework for coalescent theory because they are analytically tractable using matrix manipulations. The purpose of this review is to explain the phase-type theory and demonstrate how the theory can be applied to derive basic properties of coalescent models. These properties can then be used to obtain insight into the ancestral process, or they can be applied for statistical inference. In particular, we show the relation between classical first-step analysis of coalescent models and phase-type calculations. We also show how reward transformations in phase-type theory lead to easy calculation of covariances and correlation coefficients between e.g. tree height, tree length, external branch length, and internal branch length. Furthermore, we discuss how these quantities can be used for statistical inference based on estimating equations. Providing an alternative to previous work based on the Laplace transform, we derive likelihoods for small-size coalescent trees based on phase-type theory. Overall, our main aim is to demonstrate that phase-type distributions provide a convenient general set of tools to understand aspects of coalescent models that are otherwise difficult to derive. Throughout the review, we emphasize the versatility of the phase-type framework, which is also illustrated by our accompanying R-code. All our analyses and figures can be reproduced from code available on GitHub.

The Tree Reconstruction Game: Phylogenetic Reconstruction Using Reinforcement Learning

The computational search for the maximum-likelihood phylogenetic tree is an NP-hard problem. As such, current tree search algorithms might result in a tree that is the local optima, not the global one. Here, we introduce a paradigm shift for predicting the maximum-likelihood tree, by approximating long-term gains of likelihood rather than maximizing likelihood gain at each step of the search. Our proposed approach harnesses the power of reinforcement learning to learn an optimal search strategy, aiming at the global optimum of the search space. We show that when analyzing empirical data containing dozens of sequences, the log-likelihood improvement from the starting tree obtained by the reinforcement learning-based agent was 0.969 or higher compared to that achieved by current state-of-the-art techniques. Notably, this performance is attained without the need to perform costly likelihood optimizations apart from the training process, thus potentially allowing for an exponential increase in runtime. We exemplify this for data sets containing 15 sequences of length 18,000 bp and demonstrate that the reinforcement learning-based method is roughly three times faster than the state-of-the-art software. This study illustrates the potential of reinforcement learning in addressing the challenges of phylogenetic tree reconstruction.

Genetic structure and landscape effects on gene flow in the Neotropical lizard Norops brasiliensis (Squamata: Dactyloidae)

One key research goal of evolutionary biology is to understand the origin and maintenance of genetic variation. In the Cerrado, the South American savanna located primarily in the Central Brazilian Plateau, many hypotheses have been proposed to explain how landscape features (e.g., geographic distance, river barriers, topographic compartmentalization, and historical climatic fluctuations) have promoted genetic structure by mediating gene flow. Here, we asked whether these landscape features have influenced the genetic structure and differentiation in the lizard species Norops brasiliensis (Squamata: Dactyloidae). To achieve our goal, we used a genetic clustering analysis and estimate an effective migration surface to assess genetic structure in the focal species. Optimized isolation-by-resistance models and a simulation-based approach combined with machine learning (convolutional neural network; CNN) were then used to infer current and historical effects on population genetic structure through 12 unique landscape models. We recovered five geographically distributed populations that are separated by regions of lower-than-expected gene flow. The results of the CNN showed that geographic distance is the sole predictor of genetic variation in N. brasiliensis, and that slope, rivers, and historical climate had no discernible influence on gene flow. Our novel CNN approach was accurate (89.5%) in differentiating each landscape model. CNN and other machine learning approaches are still largely unexplored in landscape genetics studies, representing promising avenues for future research with increasingly accessible genomic datasets.

Deep Learning and Likelihood Approaches for Viral Phylogeography Converge on the Same Answers Whether the Inference Model Is Right or Wrong

Analysis of phylogenetic trees has become an essential tool in epidemiology. Likelihood-based methods fit models to phylogenies to draw inferences about the phylodynamics and history of viral transmission. However, these methods are often computationally expensive, which limits the complexity and realism of phylodynamic models and makes them ill-suited for informing policy decisions in real-time during rapidly developing outbreaks. Likelihood-free methods using deep learning are pushing the boundaries of inference beyond these constraints. In this paper, we extend, compare, and contrast a recently developed deep learning method for likelihood-free inference from trees. We trained multiple deep neural networks using phylogenies from simulated outbreaks that spread among 5 locations and found they achieve close to the same levels of accuracy as Bayesian inference under the true simulation model. We compared robustness to model misspecification of a trained neural network to that of a Bayesian method. We found that both models had comparable performance, converging on similar biases. We also implemented a method of uncertainty quantification called conformalized quantile regression that we demonstrate has similar patterns of sensitivity to model misspecification as Bayesian highest posterior density (HPD) and greatly overlap with HPDs, but have lower precision (more conservative). Finally, we trained and tested a neural network against phylogeographic data from a recent study of the SARS-Cov-2 pandemic in Europe and obtained similar estimates of region-specific epidemiological parameters and the location of the common ancestor in Europe. Along with being as accurate and robust as likelihood-based methods, our trained neural networks are on average over 3 orders of magnitude faster after training. Our results support the notion that neural networks can be trained with simulated data to accurately mimic the good and bad statistical properties of the likelihood functions of generative phylogenetic models.

DeepDive: estimating global biodiversity patterns through time using deep learning

Understanding how biodiversity has changed through time is a central goal of evolutionary biology. However, estimates of past biodiversity are challenged by the inherent incompleteness of the fossil record, even when state-of-the-art statistical methods are applied to adjust estimates while correcting for sampling biases. Here we develop an approach based on stochastic simulations of biodiversity and a deep learning model to infer richness at global or regional scales through time while incorporating spatial, temporal and taxonomic sampling variation. Our method outperforms alternative approaches across simulated datasets, especially at large spatial scales, providing robust palaeodiversity estimates under a wide range of preservation scenarios. We apply our method on two empirical datasets of different taxonomic and temporal scope: the Permian-Triassic record of marine animals and the Cenozoic evolution of proboscideans. Our estimates provide a revised quantitative assessment of two mass extinctions in the marine record and reveal rapid diversification of proboscideans following their expansion out of Africa and a >70% diversity drop in the Pleistocene.

Abiotic Stress Tolerance Boosted by Genetic Diversity in Plants

Plant breeding [...].

Computationally Efficient Demographic History Inference from Allele Frequencies with Supervised Machine Learning

Inferring past demographic history of natural populations from genomic data is of central concern in many studies across research fields. Previously, our group had developed dadi, a widely used demographic history inference method based on the allele frequency spectrum (AFS) and maximum composite-likelihood optimization. However, dadi's optimization procedure can be computationally expensive. Here, we present donni (demography optimization via neural network inference), a new inference method based on dadi that is more efficient while maintaining comparable inference accuracy. For each dadi-supported demographic model, donni simulates the expected AFS for a range of model parameters then trains a set of Mean Variance Estimation neural networks using the simulated AFS. Trained networks can then be used to instantaneously infer the model parameters from future genomic data summarized by an AFS. We demonstrate that for many demographic models, donni can infer some parameters, such as population size changes, very well and other parameters, such as migration rates and times of demographic events, fairly well. Importantly, donni provides both parameter and confidence interval estimates from input AFS with accuracy comparable to parameters inferred by dadi's likelihood optimization while bypassing its long and computationally intensive evaluation process. donni's performance demonstrates that supervised machine learning algorithms may be a promising avenue for developing more sustainable and computationally efficient demographic history inference methods.

REFORMS: Consensus-based Recommendations for Machine-learning-based Science

Machine learning (ML) methods are proliferating in scientific research. However, the adoption of these methods has been accompanied by failures of validity, reproducibility, and generalizability. These failures can hinder scientific progress, lead to false consensus around invalid claims, and undermine the credibility of ML-based science. ML methods are often applied and fail in similar ways across disciplines. Motivated by this observation, our goal is to provide clear recommendations for conducting and reporting ML-based science. Drawing from an extensive review of past literature, we present the REFORMS checklist (recommendations for machine-learning-based science). It consists of 32 questions and a paired set of guidelines. REFORMS was developed on the basis of a consensus of 19 researchers across computer science, data science, mathematics, social sciences, and biomedical sciences. REFORMS can serve as a resource for researchers when designing and implementing a study, for referees when reviewing papers, and for journals when enforcing standards for transparency and reproducibility.

Population size rescaling significantly biases outcomes of forward-in-time population genetic simulations

Simulations are an essential tool in all areas of population genetic research, used in tasks such as the validation of theoretical analysis and the study of complex evolutionary models. Forward-in-time simulations are especially flexible, allowing for various types of natural selection, complex genetic architectures, and non-Wright-Fisher dynamics. However, their intense computational requirements can be prohibitive to simulating large populations and genomes. A popular method to alleviate this burden is to scale down the population size by some scaling factor while scaling up the mutation rate, selection coefficients, and recombination rate by the same factor. However, this rescaling approach may in some cases bias simulation results. To investigate the manner and degree to which rescaling impacts simulation outcomes, we carried out simulations with different demographic histories and distributions of fitness effects using several values of the rescaling factor, Q , and compared the deviation of key outcomes (fixation times, fixation probabilities, allele frequencies, and linkage disequilibrium) between the scaled and unscaled simulations. Our results indicate that scaling introduces substantial biases to each of these measured outcomes, even at small values of Q . Moreover, the nature of these effects depends on the evolutionary model and scaling factor being examined. While increasing the scaling factor tends to increase the observed biases, this relationship is not always straightforward, thus it may be difficult to know the impact of scaling on simulation outcomes a priori. However, it appears that for most models, only a small number of replicates was needed to accurately quantify the bias produced by rescaling for a given Q . In summary, while rescaling forward-in-time simulations may be necessary in many cases, researchers should be aware of the rescaling effect's impact on simulation outcomes and consider investigating its magnitude in smaller scale simulations of the desired model(s) before selecting an appropriate value of Q .

How useful are genomic data for predicting maladaptation to future climate?

Methods using genomic information to forecast potential population maladaptation to climate change or new environments are becoming increasingly common, yet the lack of model validation poses serious hurdles toward their incorporation into management and policy. Here, we compare the validation of maladaptation estimates derived from two methods-Gradient Forests (GFoffset) and the risk of non-adaptedness (RONA)-using exome capture pool-seq data from 35 to 39 populations across three conifer taxa: two Douglas-fir varieties and jack pine. We evaluate sensitivity of these algorithms to the source of input loci (markers selected from genotype-environment associations [GEA] or those selected at random). We validate these methods against 2- and 52-year growth and mortality measured in independent transplant experiments. Overall, we find that both methods often better predict transplant performance than climatic or geographic distances. We also find that GFoffset and RONA models are surprisingly not improved using GEA candidates. Even with promising validation results, variation in model projections to future climates makes it difficult to identify the most maladapted populations using either method. Our work advances understanding of the sensitivity and applicability of these approaches, and we discuss recommendations for their future use.

A Novel Approach Utilizing Domain Adversarial Neural Networks for the Detection and Classification of Selective Sweeps

The identification and classification of selective sweeps are of great significance for improving the understanding of biological evolution and exploring opportunities for precision medicine and genetic improvement. Here, a domain adaptation sweep detection and classification (DASDC) method is presented to balance the alignment of two domains and the classification performance through a domain-adversarial neural network and its adversarial learning modules. DASDC effectively addresses the issue of mismatch between training data and real genomic data in deep learning models, leading to a significant improvement in its generalization capability, prediction robustness, and accuracy. The DASDC method demonstrates improved identification performance compared to existing methods and excels in classification performance, particularly in scenarios where there is a mismatch between application data and training data. The successful implementation of DASDC in real data of three distinct species highlights its potential as a useful tool for identifying crucial functional genes and investigating adaptive evolutionary mechanisms, particularly with the increasing availability of genomic data.

Bringing CAM photosynthesis to the table: Paving the way for resilient and productive agricultural systems in a changing climate

Modern agricultural systems are directly threatened by global climate change and the resulting freshwater crisis. A considerable challenge in the coming years will be to develop crops that can cope with the consequences of declining freshwater resources and changing temperatures. One approach to meeting this challenge may lie in our understanding of plant photosynthetic adaptations and water use efficiency. Plants from various taxa have evolved crassulacean acid metabolism (CAM), a water-conserving adaptation of photosynthetic carbon dioxide fixation that enables plants to thrive under semi-arid or seasonally drought-prone conditions. Although past research on CAM has led to a better understanding of the inner workings of plant resilience and adaptation to stress, successful introduction of this pathway into C3 or C4 plants has not been reported. The recent revolution in molecular, systems, and synthetic biology, as well as innovations in high-throughput data generation and mining, creates new opportunities to uncover the minimum genetic tool kit required to introduce CAM traits into drought-sensitive crops. Here, we propose four complementary research avenues to uncover this tool kit. First, genomes and computational methods should be used to improve understanding of the nature of variations that drive CAM evolution. Second, single-cell 'omics technologies offer the possibility for in-depth characterization of the mechanisms that trigger environmentally controlled CAM induction. Third, the rapid increase in new 'omics data enables a comprehensive, multimodal exploration of CAM. Finally, the expansion of functional genomics methods is paving the way for integration of CAM into farming systems.

Computationally efficient demographic history inference from allele frequencies with supervised machine learning

Inferring past demographic history of natural populations from genomic data is of central concern in many studies across research fields. Previously, our group had developed dadi, a widely used demographic history inference method based on the allele frequency spectrum (AFS) and maximum composite likelihood optimization. However, dadi's optimization procedure can be computationally expensive. Here, we developed donni (demography optimization via neural network inference), a new inference method based on dadi that is more efficient while maintaining comparable inference accuracy. For each dadi-supported demographic model, donni simulates the expected AFS for a range of model parameters then trains a set of Mean Variance Estimation neural networks using the simulated AFS. Trained networks can then be used to instantaneously infer the model parameters from future input data AFS. We demonstrated that for many demographic models, donni can infer some parameters, such as population size changes, very well and other parameters, such as migration rates and times of demographic events, fairly well. Importantly, donni provides both parameter and confidence interval estimates from input AFS with accuracy comparable to parameters inferred by dadi's likelihood optimization while bypassing its long and computationally intensive evaluation process. donni's performance demonstrates that supervised machine learning algorithms may be a promising avenue for developing more sustainable and computationally efficient demographic history inference methods.

IntroUNET: Identifying introgressed alleles via semantic segmentation

A growing body of evidence suggests that gene flow between closely related species is a widespread phenomenon. Alleles that introgress from one species into a close relative are typically neutral or deleterious, but sometimes confer a significant fitness advantage. Given the potential relevance to speciation and adaptation, numerous methods have therefore been devised to identify regions of the genome that have experienced introgression. Recently, supervised machine learning approaches have been shown to be highly effective for detecting introgression. One especially promising approach is to treat population genetic inference as an image classification problem, and feed an image representation of a population genetic alignment as input to a deep neural network that distinguishes among evolutionary models (i.e. introgression or no introgression). However, if we wish to investigate the full extent and fitness effects of introgression, merely identifying genomic regions in a population genetic alignment that harbor introgressed loci is insufficient-ideally we would be able to infer precisely which individuals have introgressed material and at which positions in the genome. Here we adapt a deep learning algorithm for semantic segmentation, the task of correctly identifying the type of object to which each individual pixel in an image belongs, to the task of identifying introgressed alleles. Our trained neural network is thus able to infer, for each individual in a two-population alignment, which of those individual's alleles were introgressed from the other population. We use simulated data to show that this approach is highly accurate, and that it can be readily extended to identify alleles that are introgressed from an unsampled "ghost" population, performing comparably to a supervised learning method tailored specifically to that task. Finally, we apply this method to data from Drosophila, showing that it is able to accurately recover introgressed haplotypes from real data. This analysis reveals that introgressed alleles are typically confined to lower frequencies within genic regions, suggestive of purifying selection, but are found at much higher frequencies in a region previously shown to be affected by adaptive introgression. Our method's success in recovering introgressed haplotypes in challenging real-world scenarios underscores the utility of deep learning approaches for making richer evolutionary inferences from genomic data.

IntroUNET: identifying introgressed alleles via semantic segmentation

A growing body of evidence suggests that gene flow between closely related species is a widespread phenomenon. Alleles that introgress from one species into a close relative are typically neutral or deleterious, but sometimes confer a significant fitness advantage. Given the potential relevance to speciation and adaptation, numerous methods have therefore been devised to identify regions of the genome that have experienced introgression. Recently, supervised machine learning approaches have been shown to be highly effective for detecting introgression. One especially promising approach is to treat population genetic inference as an image classification problem, and feed an image representation of a population genetic alignment as input to a deep neural network that distinguishes among evolutionary models (i.e. introgression or no introgression). However, if we wish to investigate the full extent and fitness effects of introgression, merely identifying genomic regions in a population genetic alignment that harbor introgressed loci is insufficient-ideally we would be able to infer precisely which individuals have introgressed material and at which positions in the genome. Here we adapt a deep learning algorithm for semantic segmentation, the task of correctly identifying the type of object to which each individual pixel in an image belongs, to the task of identifying introgressed alleles. Our trained neural network is thus able to infer, for each individual in a two-population alignment, which of those individual's alleles were introgressed from the other population. We use simulated data to show that this approach is highly accurate, and that it can be readily extended to identify alleles that are introgressed from an unsampled "ghost" population, performing comparably to a supervised learning method tailored specifically to that task. Finally, we apply this method to data from Drosophila, showing that it is able to accurately recover introgressed haplotypes from real data. This analysis reveals that introgressed alleles are typically confined to lower frequencies within genic regions, suggestive of purifying selection, but are found at much higher frequencies in a region previously shown to be affected by adaptive introgression. Our method's success in recovering introgressed haplotypes in challenging real-world scenarios underscores the utility of deep learning approaches for making richer evolutionary inferences from genomic data.

Assessing the predictive value of morphological traits on primary lifestyle of birds through the extreme gradient boosting algorithm

The relationship between morphological traits and ecological performance in birds is an important area of research, as it can help us to understand how birds are able to adapt and how they are affected by changes in their environment. Many studies have investigated the relationship between morphological traits and certain aspects of the performance and ecological niche of these animals. However, the relationship between morphological traits and the primary lifestyle of birds has not previously been explored. This paper aims to evaluate the predictive potential of morphological data to determine the primary lifestyle of birds through a tree-based machine learning algorithm. By doing this, it is also possible to evaluate these artificial categories that we used to split up birds and know whether they are suitable for dividing them in function of shared morphological characteristics or need a redefinition under more discriminant criteria. Supplementary dataset 1 of the AVONET project was used, which comprises the 11 morphological predictors used in this work and the classification according to the primary lifestyle for more than 95% of the existing bird species. For all morphological traits used, statistically significant univariate differences were found between primary lifestyles. The three fitted machine learning models showed high accuracy, in all cases above 78% and superior to the ones achieved through traditional approaches used as contrasts. The results obtained provide evidence that primary lifestyle can be predicted in birds based on morphological traits, as well as more insights about the relevance of functional traits for ecological modeling. This is another step forward in our mechanistic understanding of bird ecology, while exploring how birds have adapted to their environments and how they interact with their surroundings.

Unveiling recent and ongoing adaptive selection in human populations

Genome-wide scans for signals of selection have become a routine part of the analysis of population genomic variation datasets and have resulted in compelling evidence of selection during recent human evolution. This Essay spotlights methodological innovations that have enabled the detection of selection over very recent timescales, even in contemporary human populations. By harnessing large-scale genomic and phenotypic datasets, these new methods use different strategies to uncover connections between genotype, phenotype, and fitness. This Essay outlines the rationale and key findings of each strategy, discusses challenges in interpretation, and describes opportunities to improve detection and understanding of ongoing selection in human populations.

Challenges in developing cell culture media using machine learning

Microbial and mammalian cells are widely used in the food, pharmaceutical, and medical industries. Developing or optimizing culture media is essential to improve cell culture performance as a critical technology in cell culture engineering. Methodologies for media optimization have been developed to a great extent, such as the approaches of one-factor-at-a-time (OFAT) and response surface methodology (RSM). The present review introduces the emerging machine learning (ML) technology in cell culture engineering by combining high-throughput experimental technologies to develop highly efficient and effective culture media. The commonly used ML algorithms and the successful applications of employing ML in medium optimization are summarized. This review highlights the benefits of ML-assisted medium development and guides the selection of the media optimization method appropriate for various cell culture purposes.

Harnessing deep learning for population genetic inference

In population genetics, the emergence of large-scale genomic data for various species and populations has provided new opportunities to understand the evolutionary forces that drive genetic diversity using statistical inference. However, the era of population genomics presents new challenges in analysing the massive amounts of genomes and variants. Deep learning has demonstrated state-of-the-art performance for numerous applications involving large-scale data. Recently, deep learning approaches have gained popularity in population genetics; facilitated by the advent of massive genomic data sets, powerful computational hardware and complex deep learning architectures, they have been used to identify population structure, infer demographic history and investigate natural selection. Here, we introduce common deep learning architectures and provide comprehensive guidelines for implementing deep learning models for population genetic inference. We also discuss current challenges and future directions for applying deep learning in population genetics, focusing on efficiency, robustness and interpretability.

Inferring the ecological and evolutionary determinants of community genetic diversity

Understanding the relative contributions of ecological and evolutionary processes to the structuring of ecological communities is needed to improve our ability to predict how communities may respond to future changes in an increasingly human-modified world. Metabarcoding methods make it possible to gather population genetic data for all species within a community, unlocking a new axis of data to potentially unveil the origins and maintenance of biodiversity at local scales. Here, we present a new eco-evolutionary simulation model for investigating community assembly dynamics using metabarcoding data. The model makes joint predictions of species abundance, genetic variation, trait distributions and phylogenetic relationships under a wide range of parameter settings (e.g. high speciation/low dispersal or vice versa) and across a range of community states, from pristine and unmodified to heavily disturbed. We first demonstrate that parameters governing metacommunity and local community processes leave detectable signatures in simulated biodiversity data axes. Next, using a simulation-based machine learning approach we show that neutral and non-neutral models are distinguishable and that reasonable estimates of several model parameters within the local community can be obtained using only community-scale genetic data, while phylogenetic information is required to estimate those describing metacommunity dynamics. Finally, we apply the model to soil microarthropod metabarcoding data from the Troodos mountains of Cyprus, where we find that communities in widespread forest habitats are structured by neutral processes, while high-elevation and isolated habitats act as an abiotic filter generating non-neutral community structure. We implement our model within the ibiogen R package, a package dedicated to the investigation of island, and more generally community-scale, biodiversity using community-scale genetic data.

Landmarks in the history of selective sweeps

Half a century ago, a seminal article on the hitchhiking effect by Smith and Haigh inaugurated the concept of the selection signature. Selective sweeps are characterised by the rapid spread of an advantageous genetic variant through a population and hence play an important role in shaping evolution and research on genetic diversity. The process by which a beneficial allele arises and becomes fixed in a population, leading to a increase in the frequency of other linked alleles, is known as genetic hitchhiking or genetic draft. Kimura's neutral theory and hitchhiking theory are complementary, with Kimura's neutral evolution as the 'null model' and positive selection as the 'signal'. Both are widely accepted in evolution, especially with genomics enabling precise measurements. Significant advances in genomic technologies, such as next-generation sequencing, high-density SNP arrays and powerful bioinformatics tools, have made it possible to systematically investigate selection signatures in a variety of species. Although the history of selection signatures is relatively recent, progress has been made in the last two decades, owing to the increasing availability of large-scale genomic data and the development of computational methods. In this review, we embark on a journey through the history of research on selective sweeps, ranging from early theoretical work to recent empirical studies that utilise genomic data.

Data-driven approaches to improve water-use efficiency and drought resistance in crop plants

With the increasing population, there lies a pressing demand for food, feed and fibre, while the changing climatic conditions pose severe challenges for agricultural production worldwide. Water is the lifeline for crop production; thus, enhancing crop water-use efficiency (WUE) and improving drought resistance in crop varieties are crucial for overcoming these challenges. Genetically-driven improvements in yield, WUE and drought tolerance traits can buffer the worst effects of climate change on crop production in dry areas. While traditional crop breeding approaches have delivered impressive results in increasing yield, the methods remain time-consuming and are often limited by the existing allelic variation present in the germplasm. Significant advances in breeding and high-throughput omics technologies in parallel with smart agriculture practices have created avenues to dramatically speed up the process of trait improvement by leveraging the vast volumes of genomic and phenotypic data. For example, individual genome and pan-genome assemblies, along with transcriptomic, metabolomic and proteomic data from germplasm collections, characterised at phenotypic levels, could be utilised to identify marker-trait associations and superior haplotypes for crop genetic improvement. In addition, these omics approaches enable the identification of genes involved in pathways leading to the expression of a trait, thereby providing an understanding of the genetic, physiological and biochemical basis of trait variation. These data-driven gene discoveries and validation approaches are essential for crop improvement pipelines, including genomic breeding, speed breeding and gene editing. Herein, we provide an overview of prospects presented using big data-driven approaches (including artificial intelligence and machine learning) to harness new genetic gains for breeding programs and develop drought-tolerant crop varieties with favourable WUE and high-yield potential traits.

On convolutional neural networks for selection inference: Revealing the effect of preprocessing on model learning and the capacity to discover novel patterns

A central challenge in population genetics is the detection of genomic footprints of selection. As machine learning tools including convolutional neural networks (CNNs) have become more sophisticated and applied more broadly, these provide a logical next step for increasing our power to learn and detect such patterns; indeed, CNNs trained on simulated genome sequences have recently been shown to be highly effective at this task. Unlike previous approaches, which rely upon human-crafted summary statistics, these methods are able to be applied directly to raw genomic data, allowing them to potentially learn new signatures that, if well-understood, could improve the current theory surrounding selective sweeps. Towards this end, we examine a representative CNN from the literature, paring it down to the minimal complexity needed to maintain comparable performance; this low-complexity CNN allows us to directly interpret the learned evolutionary signatures. We then validate these patterns in more complex models using metrics that evaluate feature importance. Our findings reveal that preprocessing steps, which determine how the population genetic data is presented to the model, play a central role in the learned prediction method. This results in models that mimic previously-defined summary statistics; in one case, the summary statistic itself achieves similarly high accuracy. For evolutionary processes that are less well understood than selective sweeps, we hope this provides an initial framework for using CNNs in ways that go beyond simply achieving high classification performance. Instead, we propose that CNNs might be useful as tools for learning novel patterns that can translate to easy-to-implement summary statistics available to a wider community of researchers.

Domain-adaptive neural networks improve supervised machine learning based on simulated population genetic data

Investigators have recently introduced powerful methods for population genetic inference that rely on supervised machine learning from simulated data. Despite their performance advantages, these methods can fail when the simulated training data does not adequately resemble data from the real world. Here, we show that this "simulation mis-specification" problem can be framed as a "domain adaptation" problem, where a model learned from one data distribution is applied to a dataset drawn from a different distribution. By applying an established domain-adaptation technique based on a gradient reversal layer (GRL), originally introduced for image classification, we show that the effects of simulation mis-specification can be substantially mitigated. We focus our analysis on two state-of-the-art deep-learning population genetic methods-SIA, which infers positive selection from features of the ancestral recombination graph (ARG), and ReLERNN, which infers recombination rates from genotype matrices. In the case of SIA, the domain adaptive framework also compensates for ARG inference error. Using the domain-adaptive SIA (dadaSIA) model, we estimate improved selection coefficients at selected loci in the 1000 Genomes CEU population. We anticipate that domain adaptation will prove to be widely applicable in the growing use of supervised machine learning in population genetics.

Tensor Decomposition-based Feature Extraction and Classification to Detect Natural Selection from Genomic Data

Inferences of adaptive events are important for learning about traits, such as human digestion of lactose after infancy and the rapid spread of viral variants. Early efforts toward identifying footprints of natural selection from genomic data involved development of summary statistic and likelihood methods. However, such techniques are grounded in simple patterns or theoretical models that limit the complexity of settings they can explore. Due to the renaissance in artificial intelligence, machine learning methods have taken center stage in recent efforts to detect natural selection, with strategies such as convolutional neural networks applied to images of haplotypes. Yet, limitations of such techniques include estimation of large numbers of model parameters under nonconvex settings and feature identification without regard to location within an image. An alternative approach is to use tensor decomposition to extract features from multidimensional data although preserving the latent structure of the data, and to feed these features to machine learning models. Here, we adopt this framework and present a novel approach termed T-REx, which extracts features from images of haplotypes across sampled individuals using tensor decomposition, and then makes predictions from these features using classical machine learning methods. As a proof of concept, we explore the performance of T-REx on simulated neutral and selective sweep scenarios and find that it has high power and accuracy to discriminate sweeps from neutrality, robustness to common technical hurdles, and easy visualization of feature importance. Therefore, T-REx is a powerful addition to the toolkit for detecting adaptive processes from genomic data.

Power of neutrality tests for detecting natural selection

Detection of natural selection is one of the main interests in population genetics. Thus, many tests have been developed for detecting natural selection using genomic data. Although it is recognized that the utility of tests depends on several evolutionary factors, such as the timing of selection, strength of selection, frequency of selected alleles, demographic events, and initial frequency of selected allele when selection started acting (softness of selection), the relationships between such evolutionary factors and the power of tests are not yet entirely clear. In this study, we investigated the power of 4 tests: Tajiama's D, Fay and Wu's H, relative extended haplotype homozygosity (rEHH), and integrated haplotype score (iHS), under ranges of evolutionary parameters and demographic models to quantitatively expand the understanding of approaches for detecting selection. The results show that each test detects selection within a limited parameter range, and there are still wide ranges of parameters for which none of these tests work effectively. In addition, the parameter space in which each test shows the highest power overlaps the empirical results of previous research. These results indicate that our present perspective of adaptation is limited to only a part of actual adaptation.

A diagnostic model for sepsis-induced acute lung injury using a consensus machine learning approach and its therapeutic implications

BACKGROUND

A significant proportion of septic patients with acute lung injury (ALI) are recognized late due to the absence of an efficient diagnostic test, leading to the postponed treatments and consequently higher mortality. Identifying diagnostic biomarkers may improve screening to identify septic patients at high risk of ALI earlier and provide the potential effective therapeutic drugs. Machine learning represents a powerful approach for making sense of complex gene expression data to find robust ALI diagnostic biomarkers.

METHODS

The datasets were obtained from GEO and ArrayExpress databases. Following quality control and normalization, the datasets (GSE66890, GSE10474 and GSE32707) were merged as the training set, and four machine learning feature selection methods (Elastic net, SVM, random forest and XGBoost) were applied to construct the diagnostic model. The other datasets were considered as the validation sets. To further evaluate the performance and predictive value of diagnostic model, nomogram, Decision Curve Analysis (DCA) and Clinical Impact Curve (CIC) were constructed. Finally, the potential small molecular compounds interacting with selected features were explored from the CTD database.

RESULTS

The results of GSEA showed that immune response and metabolism might play an important role in the pathogenesis of sepsis-induced ALI. Then, 52 genes were identified as putative biomarkers by consensus feature selection from all four methods. Among them, 5 genes (ARHGDIB, ALDH1A1, TACR3, TREM1 and PI3) were selected by all methods and used to predict ALI diagnosis with high accuracy. The external datasets (E-MTAB-5273 and E-MTAB-5274) demonstrated that the diagnostic model had great accuracy with AUC value of 0.725 and 0.833, respectively. In addition, the nomogram, DCA and CIC showed that the diagnostic model had great performance and predictive value. Finally, the small molecular compounds (Curcumin, Tretinoin, Acetaminophen, Estradiol and Dexamethasone) were screened as the potential therapeutic agents for sepsis-induced ALI.

CONCLUSION

This consensus of multiple machine learning algorithms identified 5 genes that were able to distinguish ALI from septic patients. The diagnostic model could identify septic patients at high risk of ALI, and provide potential therapeutic targets for sepsis-induced ALI.

Leakage and the reproducibility crisis in machine-learning-based science

Machine-learning (ML) methods have gained prominence in the quantitative sciences. However, there are many known methodological pitfalls, including data leakage, in ML-based science. We systematically investigate reproducibility issues in ML-based science. Through a survey of literature in fields that have adopted ML methods, we find 17 fields where leakage has been found, collectively affecting 294 papers and, in some cases, leading to wildly overoptimistic conclusions. Based on our survey, we introduce a detailed taxonomy of eight types of leakage, ranging from textbook errors to open research problems. We propose that researchers test for each type of leakage by filling out model info sheets, which we introduce. Finally, we conduct a reproducibility study of civil war prediction, where complex ML models are believed to vastly outperform traditional statistical models such as logistic regression (LR). When the errors are corrected, complex ML models do not perform substantively better than decades-old LR models.

Domain-adaptive neural networks improve supervised machine learning based on simulated population genetic data

Investigators have recently introduced powerful methods for population genetic inference that rely on supervised machine learning from simulated data. Despite their performance advantages, these methods can fail when the simulated training data does not adequately resemble data from the real world. Here, we show that this "simulation mis-specification" problem can be framed as a "domain adaptation" problem, where a model learned from one data distribution is applied to a dataset drawn from a different distribution. By applying an established domain-adaptation technique based on a gradient reversal layer (GRL), originally introduced for image classification, we show that the effects of simulation mis-specification can be substantially mitigated. We focus our analysis on two state-of-the-art deep-learning population genetic methods-SIA, which infers positive selection from features of the ancestral recombination graph (ARG), and ReLERNN, which infers recombination rates from genotype matrices. In the case of SIA, the domain adaptive framework also compensates for ARG inference error. Using the domain-adaptive SIA (dadaSIA) model, we estimate improved selection coefficients at selected loci in the 1000 Genomes CEU population. We anticipate that domain adaptation will prove to be widely applicable in the growing use of supervised machine learning in population genetics.

Phylogenetic inference using generative adversarial networks

MOTIVATION

The application of machine learning approaches in phylogenetics has been impeded by the vast model space associated with inference. Supervised machine learning approaches require data from across this space to train models. Because of this, previous approaches have typically been limited to inferring relationships among unrooted quartets of taxa, where there are only three possible topologies. Here, we explore the potential of generative adversarial networks (GANs) to address this limitation. GANs consist of a generator and a discriminator: at each step, the generator aims to create data that is similar to real data, while the discriminator attempts to distinguish generated and real data. By using an evolutionary model as the generator, we use GANs to make evolutionary inferences. Since a new model can be considered at each iteration, heuristic searches of complex model spaces are possible. Thus, GANs offer a potential solution to the challenges of applying machine learning in phylogenetics.

RESULTS

We developed phyloGAN, a GAN that infers phylogenetic relationships among species. phyloGAN takes as input a concatenated alignment, or a set of gene alignments, and infers a phylogenetic tree either considering or ignoring gene tree heterogeneity. We explored the performance of phyloGAN for up to 15 taxa in the concatenation case and 6 taxa when considering gene tree heterogeneity. Error rates are relatively low in these simple cases. However, run times are slow and performance metrics suggest issues during training. Future work should explore novel architectures that may result in more stable and efficient GANs for phylogenetics.

AVAILABILITY AND IMPLEMENTATION

phyloGAN is available on github: https://github.com/meganlsmith/phyloGAN/.

Machine learning based lineage tree reconstruction improved with knowledge of higher level relationships between cells and genomic barcodes

Tracking cells as they divide and progress through differentiation is a fundamental step in understanding many biological processes, such as the development of organisms and progression of diseases. In this study, we investigate a machine learning approach to reconstruct lineage trees in experimental systems based on mutating synthetic genomic barcodes. We refine previously proposed methodology by embedding information of higher level relationships between cells and single-cell barcode values into a feature space. We test performance of the algorithm on shallow trees (up to 100 cells) and deep trees (up to 10 000 cells). Our proposed algorithm can improve tree reconstruction accuracy in comparison to reconstructions based on a maximum parsimony method, but this comes at a higher computational time requirement.

Efforts to Minimise the Bacterial Genome as a Free-Living Growing System

Exploring the minimal genetic requirements for cells to maintain free living is an exciting topic in biology. Multiple approaches are employed to address the question of the minimal genome. In addition to constructing the synthetic genome in the test tube, reducing the size of the wild-type genome is a practical approach for obtaining the essential genomic sequence for living cells. The well-studied Escherichia coli has been used as a model organism for genome reduction owing to its fast growth and easy manipulation. Extensive studies have reported how to reduce the bacterial genome and the collections of genomic disturbed strains acquired, which were sufficiently reviewed previously. However, the common issue of growth decrease caused by genetic disturbance remains largely unaddressed. This mini-review discusses the considerable efforts made to improve growth fitness, which was decreased due to genome reduction. The proposal and perspective are clarified for further accumulated genetic deletion to minimise the Escherichia coli genome in terms of genome reduction, experimental evolution, medium optimization, and machine learning.

Expanding the stdpopsim species catalog, and lessons learned for realistic genome simulations

Simulation is a key tool in population genetics for both methods development and empirical research, but producing simulations that recapitulate the main features of genomic datasets remains a major obstacle. Today, more realistic simulations are possible thanks to large increases in the quantity and quality of available genetic data, and the sophistication of inference and simulation software. However, implementing these simulations still requires substantial time and specialized knowledge. These challenges are especially pronounced for simulating genomes for species that are not well-studied, since it is not always clear what information is required to produce simulations with a level of realism sufficient to confidently answer a given question. The community-developed framework stdpopsim seeks to lower this barrier by facilitating the simulation of complex population genetic models using up-to-date information. The initial version of stdpopsim focused on establishing this framework using six well-characterized model species (Adrion et al., 2020). Here, we report on major improvements made in the new release of stdpopsim (version 0.2), which includes a significant expansion of the species catalog and substantial additions to simulation capabilities. Features added to improve the realism of the simulated genomes include non-crossover recombination and provision of species-specific genomic annotations. Through community-driven efforts, we expanded the number of species in the catalog more than threefold and broadened coverage across the tree of life. During the process of expanding the catalog, we have identified common sticking points and developed the best practices for setting up genome-scale simulations. We describe the input data required for generating a realistic simulation, suggest good practices for obtaining the relevant information from the literature, and discuss common pitfalls and major considerations. These improvements to stdpopsim aim to further promote the use of realistic whole-genome population genetic simulations, especially in non-model organisms, making them available, transparent, and accessible to everyone.

Dispersal inference from population genetic variation using a convolutional neural network

The geographic nature of biological dispersal shapes patterns of genetic variation over landscapes, making it possible to infer properties of dispersal from genetic variation data. Here, we present an inference tool that uses geographically distributed genotype data in combination with a convolutional neural network to estimate a critical population parameter: the mean per-generation dispersal distance. Using extensive simulation, we show that our deep learning approach is competitive with or outperforms state-of-the-art methods, particularly at small sample sizes. In addition, we evaluate varying nuisance parameters during training-including population density, demographic history, habitat size, and sampling area-and show that this strategy is effective for estimating dispersal distance when other model parameters are unknown. Whereas competing methods depend on information about local population density or accurate inference of identity-by-descent tracts, our method uses only single-nucleotide-polymorphism data and the spatial scale of sampling as input. Strikingly, and unlike other methods, our method does not use the geographic coordinates of the genotyped individuals. These features make our method, which we call "disperseNN," a potentially valuable new tool for estimating dispersal distance in nonmodel systems with whole genome data or reduced representation data. We apply disperseNN to 12 different species with publicly available data, yielding reasonable estimates for most species. Importantly, our method estimated consistently larger dispersal distances than mark-recapture calculations in the same species, which may be due to the limited geographic sampling area covered by some mark-recapture studies. Thus genetic tools like ours complement direct methods for improving our understanding of dispersal.

This population does not exist: learning the distribution of evolutionary histories with generative adversarial networks

Numerous studies over the last decade have demonstrated the utility of machine learning methods when applied to population genetic tasks. More recent studies show the potential of deep-learning methods in particular, which allow researchers to approach problems without making prior assumptions about how the data should be summarized or manipulated, instead learning their own internal representation of the data in an attempt to maximize inferential accuracy. One type of deep neural network, called Generative Adversarial Networks (GANs), can even be used to generate new data, and this approach has been used to create individual artificial human genomes free from privacy concerns. In this study, we further explore the application of GANs in population genetics by designing and training a network to learn the statistical distribution of population genetic alignments (i.e. data sets consisting of sequences from an entire population sample) under several diverse evolutionary histories-the first GAN capable of performing this task. After testing multiple different neural network architectures, we report the results of a fully differentiable Deep-Convolutional Wasserstein GAN with gradient penalty that is capable of generating artificial examples of population genetic alignments that successfully mimic key aspects of the training data, including the site-frequency spectrum, differentiation between populations, and patterns of linkage disequilibrium. We demonstrate consistent training success across various evolutionary models, including models of panmictic and subdivided populations, populations at equilibrium and experiencing changes in size, and populations experiencing either no selection or positive selection of various strengths, all without the need for extensive hyperparameter tuning. Overall, our findings highlight the ability of GANs to learn and mimic population genetic data and suggest future areas where this work can be applied in population genetics research that we discuss herein.

Localizing Post-Admixture Adaptive Variants with Object Detection on Ancestry-Painted Chromosomes

Gene flow between previously differentiated populations during the founding of an admixed or hybrid population has the potential to introduce adaptive alleles into the new population. If the adaptive allele is common in one source population, but not the other, then as the adaptive allele rises in frequency in the admixed population, genetic ancestry from the source containing the adaptive allele will increase nearby as well. Patterns of genetic ancestry have therefore been used to identify post-admixture positive selection in humans and other animals, including examples in immunity, metabolism, and animal coloration. A common method identifies regions of the genome that have local ancestry "outliers" compared with the distribution across the rest of the genome, considering each locus independently. However, we lack theoretical models for expected distributions of ancestry under various demographic scenarios, resulting in potential false positives and false negatives. Further, ancestry patterns between distant sites are often not independent. As a result, current methods tend to infer wide genomic regions containing many genes as under selection, limiting biological interpretation. Instead, we develop a deep learning object detection method applied to images generated from local ancestry-painted genomes. This approach preserves information from the surrounding genomic context and avoids potential pitfalls of user-defined summary statistics. We find the method is robust to a variety of demographic misspecifications using simulated data. Applied to human genotype data from Cabo Verde, we localize a known adaptive locus to a single narrow region compared with multiple or long windows obtained using two other ancestry-based methods.

Tensor decomposition based feature extraction and classification to detect natural selection from genomic data

Inferences of adaptive events are important for learning about traits, such as human digestion of lactose after infancy and the rapid spread of viral variants. Early efforts toward identifying footprints of natural selection from genomic data involved development of summary statistic and likelihood methods. However, such techniques are grounded in simple patterns or theoretical models that limit the complexity of settings they can explore. Due to the renaissance in artificial intelligence, machine learning methods have taken center stage in recent efforts to detect natural selection, with strategies such as convolutional neural networks applied to images of haplotypes. Yet, limitations of such techniques include estimation of large numbers of model parameters under non-convex settings and feature identification without regard to location within an image. An alternative approach is to use tensor decomposition to extract features from multidimensional data while preserving the latent structure of the data, and to feed these features to machine learning models. Here, we adopt this framework and present a novel approach termed T-REx , which extracts features from images of haplotypes across sampled individuals using tensor decomposition, and then makes predictions from these features using classical machine learning methods. As a proof of concept, we explore the performance of T-REx on simulated neutral and selective sweep scenarios and find that it has high power and accuracy to discriminate sweeps from neutrality, robustness to common technical hurdles, and easy visualization of feature importance. Therefore, T-REx is a powerful addition to the toolkit for detecting adaptive processes from genomic data.

Development of quantitative structure-retention relationship models to improve the identification of leachables in food packaging using non-targeted analysis

Quantitative structure-retention relationship (QSRR) models can be used to predict the chromatographic retention time of chemicals and facilitate the identification of unknown compounds, notably with non-targeted analysis. In this study, QSRR models were developed from the data obtained for 178 pure chemical standards and four types of analytical columns (C18, phenylhexyl, pentafluorophenyl, cyano) in liquid chromatography quadrupole time-of-flight mass spectrometry (LC-Q-TOF-MS). First, different data partitioning ratios and feature selection methods [random forest (RF) and support vector machine (SVM)] were tested to build models to predict chromatographic retention times based on 2D molecular descriptors. The internal and external performances of the non-linear (RF) and corresponding linear predictive models were systematically compared, and RF models resulted in better predictive capacities [p < 0.05, with an average PVE (proportion of variance explained) value of 0.89 ± 0.02] than linear models (0.79 ± 0.03). For each column, the resulting model was applied to identify leachables from actual plastic packaging samples. An in-depth investigation of the top 20 most intense molecular features revealed that all false-positives could be identified as outliers in the QSRR models (outside of the 95% prediction bands). Furthermore, analyzing a sample on multiple chromatographic columns and applying the associated QSRR models increased the capacity to filter false positives. Such an approach will contribute to a more effective identification of unknown or unexpected leachables in plastics (e.g. non-intended added substances), therefore refining our understanding of the chemical risks associated with food contact materials.

Deep Learning in Population Genetics

Population genetics is transitioning into a data-driven discipline thanks to the availability of large-scale genomic data and the need to study increasingly complex evolutionary scenarios. With likelihood and Bayesian approaches becoming either intractable or computationally unfeasible, machine learning, and in particular deep learning, algorithms are emerging as popular techniques for population genetic inferences. These approaches rely on algorithms that learn non-linear relationships between the input data and the model parameters being estimated through representation learning from training data sets. Deep learning algorithms currently employed in the field comprise discriminative and generative models with fully connected, convolutional, or recurrent layers. Additionally, a wide range of powerful simulators to generate training data under complex scenarios are now available. The application of deep learning to empirical data sets mostly replicates previous findings of demography reconstruction and signals of natural selection in model organisms. To showcase the feasibility of deep learning to tackle new challenges, we designed a branched architecture to detect signals of recent balancing selection from temporal haplotypic data, which exhibited good predictive performance on simulated data. Investigations on the interpretability of neural networks, their robustness to uncertain training data, and creative representation of population genetic data, will provide further opportunities for technological advancements in the field.