Aaron B. A. Shafer

Mutation rate (μ) and recombination rate (r) are fundamental population-genetic parameters, yet demographic and conservation inference in polar bears ( Ursus maritimus ) has relied on values borrowed from humans and other mammals. Known values vary by an order of magnitude across mammals, and inaccuracies in these estimates propagate into downstream analyses, such as inferences of demographic history. To address this, we used high-coverage whole-genome resequencing of eight parent-offspring trios from the Western Hudson Bay subpopulation of polar bears. We identified 329 de novo mutations and estimated a pedigree-based germline mutation rate of 1.22 x 10⁻⁸ per site per generation, which is within the mammalian range and close to values previously assumed for this species. Trio-based phasing yielded 522 crossovers across 16 meioses (32.6 per meiosis), and a pedigree-inferred recombination rate of 1.50 cM/Mb. From 44 unrelated individuals sampled from the same subpopulation, we inferred a demography-aware recombination map with a length-weighted genome-wide recombination rate of 3.03 cM/Mb (median 1.28 cM/Mb). We also identified 8,515 recombination hotspot regions covering 2.10% of the genome, accounting for 43.8% of total map length, a 20.9-fold concentration. Crossovers overlapped LD-defined hotspots 5.06-fold more often than expected by chance (p < 0.001). Nucleotide diversity was low and positively rank-correlated with recombination rate, consistent with linked selection. Genomic features explained only ∼3% of recombination rate variation, with GC content and SINE abundance the only significant predictors. These species-specific parameters provide an empirical foundation for future demographic reconstruction and population projection in this vulnerable, ice-obligate species.

Populations living in geographically isolated regions are expected to have reduced gene flow, increased genetic drift, and elevated levels of inbreeding, influencing both genetic load and possibly the distribution of structural variants. The American black bear (Ursus americanus) is widely found across Ontario, Canada, where populations inhabit diverse landscapes ranging from relatively intact contiguous forested landscapes to isolated regions such as the Saugeen Peninsula. Here, we investigated the population genomic history across five Ontario black bear populations using a combination of short-read (SR) and long-read (LR) sequencing data. We compared population genetic metrics including nucleotide diversity (π), Tajimas D, runs of homozygosity (ROH), and genome-wide structural variants (SVs) to assess population health and identify signatures of isolation. Our results showed that the isolated Saugeen Peninsula population is genetically distinct and exhibited reduced nucleotide diversity, greater ROH and SV abundance compared with mainland populations. The population also showed elevated genetic load in all genomic features, suggesting increased exposure of deleterious variation that are expressed in ROH and population specific SVs. We further discuss the integration of ROH, SVs and genetic load in conservation assessments.

Ecotypes offer a window into the earliest stages of adaptive divergence, before reproductive isolation is complete. In eastern Canada, migratory and boreal caribou (Rangifer tarandus caribou) represent a striking example of recent ecological divergence, differing primarily in movement strategy despite ongoing gene flow. We used low- and high-coverage whole-genome sequencing from migratory and boreal populations in Quebec and Labrador to characterize the timing, demographic history, and genomic signatures of this divergence. Demographic modeling supported a secondary contact scenario, with an estimated divergence time roughly 7,000–11,700 years ago, coinciding with the last deglaciation period. Despite weak overall population structure (genome-wide FST = 0.02) and evidence of continued gene flow, we identified distinct genomic peaks of differentiation, nearly half of which coincided with divergent haplotype length (XP-nSL), a pattern consistent with recent, incomplete soft selective sweeps rather than classic hard sweeps. Combining differentiation-, diversity-, and haplotype-based statistics proved essential for detecting this subtle selection signature. Candidate genes under selection point to a polygenic architecture spanning bone and connective tissue development (including antler-related pathways), mitochondrial energy metabolism, and other traits plausibly linked to the ecological differences between ecotypes. Together, our results suggest that caribou ecotypes are not fixed evolutionary endpoints but a dichotomy that can re-emerge from standing genetic variation as opportunity arises.

Population age structure has important consequences for population dynamics and is a key parameter for wildlife demography and management. Epigenetic clocks have emerged as accurate molecular tools for estimating chronological age. Ungulates are among the most intensively managed wildlife worldwide, yet epigenetic clocks have been developed for only two cervid species. Here, we develop a novel epigenetic clock for mule deer ( Odocoileus hemionus ) and a multi-species clock for the Odocoileus genus, and use them to investigate variation in age-associated DNA methylation across the mule deer lifespan. The mule deer clock achieved accuracy comparable to recently developed mammalian epigenetic clocks (r ≈ 0.9; MAE ≈ 5% of species lifespan). The multi-species Odocoileus clock showed lower predictive performance (r = 0.72; MA = 0.782 years), but its broader taxonomic applicability may be particularly valuable where Odocoileus species occur sympatrically and hybridize. Cross-sectional epigenetic age estimates supported nonlinear age-associated methylation dynamics across the mule deer lifespan, whereas longitudinal changes in epigenetic age indicated that clocks optimized for cross-sectional age estimation may lack sufficient resolution to quantify short-term within-individual changes. Overall, our results highlight the value of integrating clock-level, CpG-level, and longitudinal analyses to investigate aging in wild populations. We also identify strongly age-associated loci that provide candidates for developing targeted, scalable, and lower-cost assays.

Using fluorimetry to probe post-mortem biomolecular changes in soft tissue of pig and human remains in a northern continental climate.

Understanding the natural process of decomposition of human remains in the outdoors is relevant to death investigations, particularly in cold climates where empirical data is limited. Pigs (Sus scrofa) are usually used as proxies for human decomposition due to their availability and their comparable body anatomy. This study explored decomposition patterns of pig carcasses across winter in a humid continental climate (Dfb). We evaluated decomposition patterns and total body scoring (TBS) systems across accumulated degree days (ADDs) to assess their suitability under cold Dfb conditions. Winter decomposition was prolonged, characterized by extended fresh stages and eventual mummification, while two spring carcasses used as references followed a more rapid and typical trajectory. Snow cover and soft tissue sampling influenced decomposition, underscoring the importance of accounting for environmental and procedural variables. Comparisons of three TBS systems revealed discrepancies between listed criteria and observed patterns, particularly in early stages, highlighting important considerations in their application under cold conditions. Mixed-effects models demonstrated strong positive relationships between decomposition metrics and the post-mortem interval (PMI), with generalized additive models (GAM) providing the highest R 2 and minimal inter-carcass variability. These findings emphasize the need to adapt existing methods or develop new standards for more accurate PMI estimation in cold climates.

Post-mortem interval (PMI) estimation is an important element in medico-legal death investigations, as reliable estimates can establish timelines and generate investigative leads. Various analytical techniques are used to monitor biomolecular degradation throughout decomposition, with promising results obtained from fluorescence spectroscopy. In this work, we analyzed the fluorescent profile of cell lysates from 96 biopsy punches collected from 14 human cadavers at a human taphonomic facility in Québec, Canada, with PMIs ranging from 2 to 1098 days. Excitation-emission matrices (EEMs) were acquired for each lysate and deconvoluted using parallel factor analysis for fluorophore identification. We analyzed both composite and tissue-specific (muscle, adipose, and skin) samples. Overall, tryptophan fluorescence decreased with increasing PMI, whereas fluorescence between 400-500 nm, attributed to fluorescent oxidation products, increased. Fluorescence in skin and muscle was dominated by tryptophan at early PMIs, whereas adipose exhibited minimal signal. As PMI increased, EEM fluorescence became progressively more homogeneous across tissue types, indicating that tissue-specific contributions decrease as decomposition advances. Six machine learning models were built from the composite samples to predict PMI, and XGBoost performed best, with an R ²  of 0.88 and RMSE of 127.2 days, corresponding to an error of 11.6% relative to the maximum PMI. Our findings demonstrate that fluorescence analysis of cell lysates is practical and yields accurate PMI estimates, provides important insight into decomposition-related biomolecular degradation, and can also be readily integrated into DNA/RNA co-extractions.

Microbial succession can be used to estimate the postmortem interval (PMI); however, the impact of spatial variability within the cadaver decomposition island (CDI) is not well understood. This study examined spatial variation in necrobiome communities where soil samples were collected over time and across spatial locations from the CDIs of two human body donors. Microbial communities were characterized using 16S rRNA sequencing and statistical modelling of variation and PMI were conducted. Necrobiome community metrics showed no significant differences across anatomical sampling sites within the CDI at a single timepoint. Temporal modelling identified 11 taxa with significant relationships to PMI in one donor, with spatial sampling having a minimal impact on the PMI relationships. Non-linear approaches also identified taxa with likely PMI signals in the second donor. These findings demonstrate that opportunistic sampling can capture robust linear and non-linear PMI signals in later decomposition stages.