Machine‐Learning Models Trained on Genomic Summary Statistics Assess Extinction Risk and Recovery Potential of Bottlenecked Populations
5 affiliations
- 1School of Environmental Sciences University of East Anglia, Norwich Research Park Norwich UK
- 2Mansfield College University of Oxford Oxford UK
- 3School of Computing Sciences University of East Anglia, Norwich Research Park Norwich UK
- 4Durrell Institute of Conservation and Ecology (DICE), School of Natural Sciences University of Kent Canterbury UK
- 5Globe Institute University of Copenhagen Copenhagen Denmark
Abstract
Forecasting extinction risk from genomic data is challenging because of interactions among environment, demography, and genomic erosion. Here, we use known outcomes from forward simulations as a proof‐of‐concept study to test whether machine‐learning (ML) models can discriminate relative extinction risk and recovery potential from genomic summary statistics, and identify the most informative predictors. Using individual‐based, forward‐in‐time simulations, we generated genomic and fitness data for 11,200 populations across 112 demographic parameter combinations (ancestral population size: 100–10,000, bottleneck size: 4–50, fecundity: 8–64). We evaluated populations at three time points relative to bottleneck onset: 100 years before, 10 years before and at onset. We then assessed their subsequent fate up to 200 years after the end of the bottleneck, classifying populations as extinct or recovered. From 10 sampled individuals per population, we derived 738 predictors spanning diversity, heterozygosity, genetic/realised/masked load, and ROH‐based inbreeding (FROH). Unsupervised analyses primarily captured historical demography and did not separate outcomes. In contrast, supervised models achieved high performance at bottleneck onset, correctly predicting population fate (i.e., future extinction or recovery) from genomics alone with ≈86%–87% accuracy. However, extinction was harder to predict correctly than recovery because simulated extinctions were rarer. Bottleneck size and fecundity provided strong baseline predictability, with genomics adding modest but complementary information. The upper‐quantile ROH/FROH summaries were consistently more informative than mean diversity or mean load. A complementary single‐genome analysis showed that reference‐genome‐style summaries retain useful signal when demographic information is sparse. Together, these results highlight the promise of ML‐informed extinction‐risk assessment to translate genomic indicators into more actionable, early‐warning signals for biodiversity conservation.
Related
- Evolutionary rescue model informs strategies for driving cancer cell populations to extinction
- Bottlenecked but Booming: Strong Inbreeding in a Highly Invasive Freshwater Mussel
- Osteohistological evidence for terrestrial adaptation and an extended growth period in mekosuchine crocodylians and its potential role in their extinction
- Fear learning–induced brain dynamics predict individual extinction memory expression following transcranial magnetic stimulation
- Deep learning for mass extinction detection on fossilized phylogenies: power, limitations and lessons for simulation-based birth–death inference
BibTeX
@article{winder2026machinelearning,
title = {{Machine‐Learning Models Trained on Genomic Summary Statistics Assess Extinction Risk and Recovery Potential of Bottlenecked Populations}},
author = {Johanna C. Winder and Jack N. Marsters and Thomas Birley and Jasmyn Gooding and Taoyang Wu and Jim J. Groombridge and Hernán E. Morales and Cock vanOosterhout},
journal = {Evolutionary Applications},
year = {2026},
month = oct,
doi = {10.1111/eva.70331},
url = {https://doi.org/10.1111/eva.70331},
note = {CC BY},
}
RIS
TY - JOUR
TI - Machine‐Learning Models Trained on Genomic Summary Statistics Assess Extinction Risk and Recovery Potential of Bottlenecked Populations
AU - Johanna C. Winder
AU - Jack N. Marsters
AU - Thomas Birley
AU - Jasmyn Gooding
AU - Taoyang Wu
AU - Jim J. Groombridge
AU - Hernán E. Morales
AU - Cock vanOosterhout
PY - 2026
DA - 2026/10/28
JO - Evolutionary Applications
DO - 10.1111/eva.70331
UR - https://doi.org/10.1111/eva.70331
AB - Forecasting extinction risk from genomic data is challenging because of interactions among environment, demography, and genomic erosion. Here, we use known outcomes from forward simulations as a proof‐of‐concept study to test whether machine‐learning (ML) models can discriminate relative extinction risk and recovery potential from genomic summary statistics, and identify the most informative predictors. Using individual‐based, forward‐in‐time simulations, we generated genomic and fitness data for 11,200 populations across 112 demographic parameter combinations (ancestral population size: 100–10,000, bottleneck size: 4–50, fecundity: 8–64). We evaluated populations at three time points relative to bottleneck onset: 100 years before, 10 years before and at onset. We then assessed their subsequent fate up to 200 years after the end of the bottleneck, classifying populations as extinct or recovered. From 10 sampled individuals per population, we derived 738 predictors spanning diversity, heterozygosity, genetic/realised/masked load, and ROH‐based inbreeding (FROH). Unsupervised analyses primarily captured historical demography and did not separate outcomes. In contrast, supervised models achieved high performance at bottleneck onset, correctly predicting population fate (i.e., future extinction or recovery) from genomics alone with ≈86%–87% accuracy. However, extinction was harder to predict correctly than recovery because simulated extinctions were rarer. Bottleneck size and fecundity provided strong baseline predictability, with genomics adding modest but complementary information. The upper‐quantile ROH/FROH summaries were consistently more informative than mean diversity or mean load. A complementary single‐genome analysis showed that reference‐genome‐style summaries retain useful signal when demographic information is sparse. Together, these results highlight the promise of ML‐informed extinction‐risk assessment to translate genomic indicators into more actionable, early‐warning signals for biodiversity conservation.
SN - 1752-4571
ER -