Genetics Selection Evolution

BackgroundGenomic prediction in livestock is predominantly based on additive models, even though dominance and other non-additive effects can contribute appreciably to phenotypic variance for fitness and fertility traits. Bayesian mixture models, such as Bayes R, have proven effective for modelling sparse, heterogeneous additive SNP effects, but most implementations do not explicitly accommodate dominance. In this study, we extended BayesR3 to jointly model additive and dominance marker effects within a unified Bayesian mixture framework, denoted BayesR3AD, and used this method to estimate additive and dominance effects for fertility and cow survival (longevity) in Holstein cattle. ResultsUsing real Holstein genotypes (227,942 animals, 74,626 SNPs), we simulated phenotypes with additive and dominance effects. When dominance was present in the simulated data, BayesR3AD improved prediction accuracy of genetic values by +0.1011 (0.6144 vs 0.5133; {approx}19.7% relative) compared with the additive-only BayesR3 model and recovered additive and dominance variance components without bias. Under purely additive simulations, dominance mixture components were effectively empty, confirming that the extended model shrinks unnecessary dominance effects toward zero. In real fertility data, including calving interval (63,378 records) and survival (68,514 records), BayesR3AD estimated small dominance variance ({approx}1-3% of total genetic variance). The model highlighted a very large additive loci at 57.82 Mb on BTA18 for both calving interval and survival. concordant with previous GWAS studies of Holstein fertility. Additionally, a large dominance effect was found at 44.37 Mb on BTA18 for calving interval implicating a heterozygote advantage that increases fertility. ConclusionsBayesR3AD provides a practical extension of BayesR3 that captures both additive and dominance contributions to genomic prediction. The method is robust, reverting effectively to the additive model when dominance is absent, while delivering accurate variance decomposition, and potential gains in prediction accuracy when dominance is present. Application to Holstein fertility traits demonstrates that dominance can be detected and quantified without compromising additive inference, supporting improved prediction of total genetic merit. While validated in cattle, BayesR3AD can be directly applied to other species to better model and predict traits related to fitness.

Data from 80,713 first-calving cows (1984 to1989) of the Holstein, Mambi, and Siboney breeds, belonging to seven large dairy enterprises in Cuba and progenies of 1,297 sires, were analyzed. For each cow, the average across all lactations for at least 14 years after first calving was defined as individual productivity (PI), and the corresponding lifetime sum as accumulated productivity (PA); both traits were. Two genetic models were fitted: a classical Animal Model (M1) and a Sire maternal grandsire model (Sire MGS; M2), aimed at partitioning additive genetic variance into paternal and maternal-line components. Heritability estimates under model M1 were moderate (h2 {approx} 0.135 to 0.140), whereas M2 yielded higher values (h2 {approx} 0.158 to 0.170), reflecting increased additive variance due to a better connectedness across herds. Using estimated breeding values (EBV) for PI and PA, a global cow merit index (H1) was defined under M1. Under M2, a parental index (IM2) combining four standardized predictors (paternal and maternal-grandsire EBV for PI and PA) was constructed. Multiple regression of H1 on IM2 showed that the paternal and maternal-grandsire paths accounted for 73% and 27% of the variation, respectively, indicating a non-negligible maternal-line contribution. Model M2 provided the best overall fit according to information criteria and cross validation using two independent subsamples and the full population yielded correlations of 0.870 to 0.881, demonstrating strong predictive ability and stability of IM2 rankings. These results support the Sire MGS model as a structural extension of the Animal Model for breeding programs targeting lifetime productivity in tropical dairy cattle.

Breeding programs are essential in aquaculture, improving economically and environmentally important traits. In aquaculture systems, animals are raised in large groups, where social interactions are frequent and can influence individual performance. In these circumstances, indirect genetic effects can play an important role in the response to selection, and consequently, their effects on selection outcomes must be analyzed. This study aimed to evaluate the implications of heterogeneous social interaction effects on fish breeding programs using stochastic simulations. We simulated a fish breeding program with 2000 selection candidates from 1000 families formed by a partial mating design of 100 males and 100 females. Social interactions were simulated, affected by the target phenotype and two latent-personality traits. We investigated how genetic gains and phenotypic variances are affected by the magnitude and direction of social interaction effects on the target phenotype, different selection strategies, and the genetic correlations between the target phenotype and personality traits. Our results showed that increased social interaction effects lead to greater phenotypic variability in the target trait. Under mass selection, the genetic means of personality traits change, and these changes depend on the strength and direction of genetic correlations between the focal and personality traits. Conversely, group selection did not increase phenotypic variability but reduced genetic gain for the focal trait compared to mass selection. Moreover, group selection did not alter the genetic means of personality traits. However, this approach increased the rate of inbreeding per generation, which could be mitigated by optimizing the number of families per group.

BackgroundGenomic selection (GS) has revolutionized animal breeding, spanning livestock sectors such as pigs and cattle to aquatic species like fish and shrimp. However, its broader application across these industries is often constrained by high genotyping costs and reduced predictive reliability across divergent populations or generations. Developing cost-effective, biologically informed genotyping strategies to overcome these limitations remains a critical goal in animal agriculture. Epigenetic annotations, particularly histone modifications, provide direct functional insights into regulatory elements underlying complex trait variation and represent a promising but underexplored resource for marker prioritization. ResultsHere, using the Pacific white shrimp (Litopenaeus vannamei) as a model organism, we conducted a proof-of-concept study integrating resequencing and phenotypic data from 972 individuals. We generated high-resolution epigenomic maps by profiling four histone marks (H3K4me1, H3K4me3, H3K27me3, and H3K27ac) across multiple embryonic stages and adult muscle tissue using CUT&Tag. These functional annotations were then leveraged to prioritize single nucleotide polymorphism (SNP) subsets for genomic prediction. Among the tested strategies, SNPs located in the muscle-specific bivalent promoter/enhancer (E6) state--characterized by the co-occurrence of active and repressive marks--consistently maximized prediction accuracy under the BayesA model. Notably, even at a moderate density (15k), E6-derived SNPs achieved prediction accuracies exceeding those obtained using substantially larger genome-wide SNP sets. Most importantly, in a challenging cross-population validation using an independent strain, the E6-derived SNP subset significantly improved prediction accuracy by 47.6% (increasing from 0.21 {+/-} 0.05 to 0.31 {+/-} 0.04, p < 0.05) compared to random subsets at equivalent density. ConclusionsThese results demonstrate that epigenetic annotation-guided SNP prioritization provides a biologically informed and cost-effective strategy to enhance genomic prediction accuracy and stability. This framework is broadly transferable across species and offers a practical strategy for designing low-density genotyping panels that reduce costs while maintaining reliable selection outcomes in large-scale breeding programs.

BackgroundGene expression levels are affected by genetics and environmental effects. However, quantification of the influence of genetics and environmental effects on gene expression remains limited, especially in farm animals. Here, the relative influence of genetic and heat-related environmental variations on gene expression levels was investigated in pigs, using a backcross herd of diverse heat adaptation levels. Backcross animals were raised in either a tropical or temperate environment. Animals raised in temperate environment were subjected to an experimental heat stress at the end of their growth. ResultsWe identified 1,967 differentially expressed genes (DEGs) between pigs raised in the tropical (n = 181) and temperate (n = 180) facilities, and 472 DEGs throughout a 3 weeks experimental heat stress. Transcriptome-wide association (TWAS) study identified 139 associations between gene expression levels and thermoregulation/production traits. We detected 6,014 expression quantitative trait loci (eQTLs) associated with the expression level of 3,297 genes. Genetic variance was estimated to explain 36.3% of gene expression variance on average, and was the main source of variance for 27.7% of transcripts. Most eQTLs found are located in proximal regions (cis-eQTLs) and few within distal regions (trans-eQTLs) to their assigned genes. A trans-eQTL hotspot highlighted a hematopoietic mechanism driven by GPATCH8. An integration of GWAS and TWAS pointed to TMCO1 and ZNF184 as candidate genes for backfat thickness. ConclusionsThis study provides a better understanding of the impact of climate, heat stress and genetic influences on the pig whole blood transcriptome.

The measurement of inbreeding has gained significance across diverse fields, including population and conservation genetics, agricultural genetics, breeding programs for animals and plants, and wildlife management. This is due to the fact that inbreeding leads to increased homozygosity and results in lower genetic diversity, rendering populations more vulnerable to environmental changes, diseases, and other stressors. High or mid-coverage whole genome sequencing (WGS) has been widely used for inbreeding estimation, but it is resource-intensive. We aimed to investigate the use of ultra low-coverage whole genome sequencing (ulcWGS) as a cost-effective alternative for inbreeding analysis. Domestic dogs were used for our study as their extensive breeding histories lead to populations with a wide range of inbreeding levels. We constructed a multi-breed reference panel from high-coverage WGS samples. Inbreeding in independent ulcWGS samples was then estimated using runs of homozygosity (RoH) and inbreeding coefficients (F). We modeled the relationship between these measures and sequencing depth using nonlinear regression, to generate inbreeding estimates relative to sequencing depth. Resulting relative RoH and F measurements were significantly correlated, with purebred dogs exhibiting more runs of homozygosity and higher inbreeding coefficients compared to mixed-breed dogs. Our findings demonstrate that ulcWGS can provide reliable and economical estimations of inbreeding, expanding accessibility to genetic monitoring.

BackgroundRecombination of parental haplotypes is a fundamental biological process that ensures proper segregation of homologous chromosomes and creates new combinations of alleles during meiosis. Crossover events are typically detected from large-scale pedigree-based genetic studies or linkage disequilibrium-based recombination maps, although these are generally limited to SNPs. Increasing amounts of long read sequencing and haplotype-resolved assemblies offer an alternative approach to examining recombination events at basepair resolution, albeit with much smaller sample sizes. ResultsHere, we analyse five high-quality genome assemblies from the Simmental cattle breed, including a newly assembled triobinned HiFi assembly of an Eringer x Simmental cross (N50 of 77 Mb and a k-mer quality value of 55.3). We integrate the five assemblies, of which two originate from maternal half-siblings, into a reference-free Simmental-specific pangenome. By considering path similarities in the pangenome, we were able to identify putative crossover events in the haplotypes of the half-siblings, as well as a greater number of events relative to the cousin due to an additional degree of generational separation. We validated the pangenome approach with phased SNPs called from linear alignments of maternal short read sequencing, with 23 of 30 chromosomes having the same recombination predictions. We identified 5 and 16.7 Mb of non-reference insertion sequences respectively shared or private to the half-siblings, enabling testing for recombination events beyond only SNP markers. We also identified four differentially methylated CpG clusters from the 5mC signal of HiFi reads which allowed us to narrow the window containing the putative recombination event from 35 to 20 Mb within the longest run of homozygosity. ConclusionStructural variants and methylation information identified from long read sequencing and genome assemblies may help identify recombination events in regions beyond those typically called from SNPs. Furthermore, while existing long read-based methylation calls can be noisy and report unrealistic intermediate methylation levels, 5mC methylation appears to be a promising avenue for distinguishing haplotypes in the absence of genomic variation.

Structural variants (SVs) are most effectively identified using long-read (LR) sequencing. However, long-read (LR) data remain limited, and sequenced samples often lack associated phenotypic information. To overcome this limitation, we combined pangenome-based (variation graph) and imputation approaches to enable large-scale SV association studies in the three main French dairy cattle breeds. A variation graph was constructed using 69,892 deletions, 89,900 insertions, and 17,402 duplications detected in 176 LR samples. We subsequently genotyped 939 samples for each SV in the panel by realigning their short read (SR) sequences to the graph. Validation analyses showed high genotype concordance rates for deletions (0.79) and insertions (0.79); however, the rates for duplications were low (0.14), leading to their exclusion from this study. Retained SVs were combined with single nucleotide variants (SNVs) and served as sequence-level imputation reference panel. From the SNP genotyping array data, we imputed SVs and SNVs for 11,902 Holstein, 3,753 Montbeliarde, and 3,053 Normande bulls. After quality control, more than 14 million SNVs and 40 thousand SVs were retained for within-breed genome-wide association analyses (GWAS) with daughter yield deviations for 13 traits related to milk production, udder health, fertility, and stature. The results of the GWAS demonstrated genetic architectures aligning with earlier discoveries and uncovered thirty-six unique significant associations between structural variants and traits. Conditional analysis revealed that ten of these SVs were the primary variants in the quantitative trait loci related to fat content, protein content, and stature.

BackgroundAlthough the biological mechanism for heterosis has been debated for a long time, heterosis is widely utilized to increase the global productivity of crops and livestock. Recently, the mechanism has been well characterized in crops and livestock with a male-heterogametic XY system due to genomic assembly advancements, especially the availability of haploid genomes. However, the biological mechanism for heterosis remains unclear in poultry possessing the female-heterogametic ZW system. ResultsHere, we assembled chromosome-level diploid and haploid genomes of the Muscovy duck. We developed an efficient and cost-effective method to assemble 12 variation graph-haploid Muscovy duck genomes from three full-sibling pairs with high quality using short-read Illumina sequences. We further characterized genetic, expression and regulatory patterns of parental alleles at multiple scales. We found that maternal haploid genomes generally had more open chromatin organization and higher accessibility, and higher levels of gene expression, while showing similar DNA methylation levels when compared to paternal haploid genomes. In contrast, the female paternal Z chromosome showed the most, and the male paternal Z chromosome presented more, relaxed chromatin organization and chromatin accessibility, and gene expression compared to the male maternal Z chromosome. Thus, the ZW system largely relies on compensation and balance to regulate gene expression on the sex Z chromosome. Moreover, we identified non-Mendelian regions covering 0.26% of the genome ([~]3.18 Mb). These regions contained lower gene density, GC content, and repeat sequence frequency, but were enriched for DNA motifs bound by transcription factors, likely leading to a compacted chromatin structure and lower chromatin accessibility. ConclusionsOur work here provides a comprehensive profile of parental alleles genetic, expression and regulatory patterns in the female-heterogametic ZW system, and might be useful for the utilization of heterosis in poultry.

Rainbow trout is affected by a broad range of pathogens causing large economic losses and animal welfare concerns. Marker-assisted selection can significantly enhance resistance to pathogens in a few generations, and to this end many studies have focused on identifying quantitative trait loci (QTLs) for resistance traits. The integration of accumulated genetic resources provides an opportunity to uncover important genetic variation and candidate genes crucially involved in rainbow trout immunity. Here, we present a comprehensive meta-QTL (MQTL) analysis based on the integration of 145 QTLs related to pathogen resistance. These QTLs were refined into 26 MQTLs, of which 15 were validated by genome-wide association studies (GWAS). The average confidence interval (CI) of these MQTLs was reduced by 2.03-fold compared to the initial QTL, improving mapping precision. Integration of GWAS results revealed regions along the rainbow trout genome pivotal for pathogen resistance, and a major region in chromosome 3, which could be used in marker-assisted selection. Further, among the validated MQTLs we identified a subset of high-confidence MQTLs, based on those supported by at least three initial QTL from more than two independent studies, with a percentage of variance explained greater than 8% and a LOD score higher than three. Gene annotation identified 11 unique candidate genes within these high-confidence MQTLs involved in immune pathways, encoding proteins involved in the regulation of immune responses, signalling pathways, receptor activity, and direct immune effector production. The MQTLs and candidate genes identified are valuable resources for advancing molecular breeding and unravelling the genetic basis of pathogen resistance in rainbow trout.

Linkage disequilibrium (LD) and haplotype block structure govern the resolution and utility of genomic selection, marker-assisted selection, and genome-wide association studies (GWAS) in livestock. We performed a comprehensive genome-wide characterization of LD decay, haplotype block architecture, and population diversity across all 24 autosomes in Nili-Ravi buffalo (Bubalus bubalis; n = 85), using 43,543 post-quality-control SNPs. Mean genome-wide r2 was 0.124 (median 0.074) and mean D was 0.540 (median 0.481), with LD half-decay at {approx}70 kb. A total of 133 haplotype blocks encompassing 721 SNPs were identified (Gabriel et al., 2002). Haploview analysis of nine chromosomes harbouring bTB resistance candidate genes revealed contrasting selection signatures: directional selection at innate immune loci (IFNG, TLR1; H < 0.55) versus balancing selection at adaptive immune loci (BoLA-DRB3, SP110; H > 1.0). Critically, BBU15 Block 3 (28.6 kb; OR52E5/NCR1 locus, 47.16 Mb) showed a genome-wide significant integrated haplotype score (iHS; -log1 0 p = 5.408), directly co-localising with the published bTB susceptibility QTL (Bermingham et al., 2014). The TAA haplotype (frequency 53.3%) at this block represents a candidate resistance-associated haplotype for marker-assisted selection. These findings provide essential parameters for SNP panel design and bTB resistance breeding in South Asian buffalo.

BackgroundUnderstanding the genetic architecture of domestic dogs provides unique insights into the processes of domestication, breed formation, and the genetic basis of complex traits and diseases. Dog populations, characterized by their diverse morphologies and behaviors, also exhibit extensive evidence of historical and ongoing admixture. This widespread mixing, driven by both natural migration and selective breeding practices, has profoundly shaped the genomic landscape of modern dog breeds. Though global admixture has been extensively estimated in human population studies, where the number of subgroups is typically limited, there has been more limited analysis in canines, where there may be dozens of ancestral groups, or breeds. ResultsHere we present a procedure for estimating global admixture in dogs from whole genome sequence data using SCOPE. We created a reference population of 65 dog breeds that included 349 individuals, from which we determined breed-informative SNPs. We demonstrate that SCOPE can accurately infer breed composition in both simulated and real admixed samples, even at low sequencing depths. We also characterized the genetic similarity between our reference dog breeds and recovered previously reported relationships. ConclusionThis approach allows us to identify the strength of the genetic signature of breeds and place error bounds on admixture estimates. It also provides evidence that admixture can be accurately inferred in subjects that may originate from multiple ancestral populations.

Plumage colour in domestic geese is an important economic trait and a selection target since the early days of domestication. In European domestic geese of greylag goose (Anser anser) origin, white plumage colour is known to be caused by two independent loci, one causing white spotting and one causing sex-linked dilution, together producing white plumage. Strong candidate mutations have been identified upstream of the EDNRB2/LOC106047519 gene (endothelin receptor B-like) and within the sex-linked MLANA gene (melan-A). To confirm these candidate mutations, we genotyped differently coloured European domestic goose breeds, wild greylag geese, Chinese domestic geese (derived from swan goose A. cygnoid) and European and Chinese domestic geese crossbreeds. One base pair deletion in the MLANA gene (NW_013185876.1: g.950,868 C > -) was confirmed to cause sex-linked dilution, and thus autosexing (almost white gander and goose diluted grey). However, mutation upstream of EDNRB2/LOC106047519 (NW_013185915.1: g. 775,151 G > T) was not the causative mutation for saddleback pattern but strongly linked to it in European domestic geese. We sequenced the EDNRB2 gene and coding sequence of a neighbouring VAMP7 gene (vesicle-associated membrane protein 7) but found no genetic variaion linked to colour. Additionally, we sequenced the coding sequence of TYRP1 (tyrosinase related protein 1), a candidate gene for buff colouration, but no variation linked to colour was found. Further, we genotyped a 14-bp insertion in exon 3 of the EDNRB2 gene, known to be causative of the white phenotype in the Chinese domestic goose, and identified it in one European domestic goose individual.

Detecting signals of polygenic adaptation remains a significant challenge in population genomics, as traditional methods often struggle to identify the associated subtle, multi-locus allele-frequency shifts. Here, we introduced and tested several novel approaches combining machine learning techniques with traditional statistical tests to detect polygenic adaptation patterns in time-series of allele frequency changes from whole genome data. We implemented a Naive Bayesian Classifier (NBC) and One-Class Support Vector Machines (OCSVM), and compared their performance against the classical Fishers Exact Test (FET). Furthermore, we combined machine learning and statistical models (OCSVM-FET and NBC-FET), resulting in 5 competing approaches. Using a simulated data set based on empirical evolve-and-resequencing Chironomus riparius genomic data, we evaluated methods across evolutionary scenarios, varying in generations, selection strength and numbers of loci under selection. Our results demonstrate that the combined OCSVM-FET approach consistently outperformed competing methods, achieving the lowest false positive rate, highest area under the curve, and high accuracy. The performance peak aligned with what we term the late dynamic phase of adaptation--the period after initial selection has occurred but before fixation--highlighting the methods sensitivity to ongoing selective processes and thus its value for experimental approaches. Furthermore, we emphasize the critical role of parameter tuning, balancing biological assumptions with methodological rigor. Our approach offers a powerful tool for detecting polygenic adaptation from time series, e.g. pool sequencing data from evolve-and-resequence experiments.

BackgroundFood intake is a complex trait in living organisms, where the genetics of food intake have been widely studied in humans, mice, Drosophila, cattle, pigs, chicken, and fish. In dairy cattle, intake of feed is highly linked to individuals energy balance, health, production, efficiency, and the environmental footprint of the individual to the society. Recent studies have provided solid evidence of the genetic variation of feed intake (FI) in dairy cattle population, but the genetic basis and molecular mechanism of dairy feed intake is still far from clear especially considering the lactation cycles of dairy cattle. This study aims to integrate stage-dependent genome-wide association (GWA) analyses, regional heritability mapping (RHM), and RNA-seq gene expression analyses to identify temporal functional variants associated with cattle dry matter intake (DMI) across multiple stages in lactation cycles. A total of 750,000 daily DMI records from 7,500 lactations of 2,300 cows were available with animals genotype and pedigree information. Total RNA-seq from blood were generated for 121 individuals in this population from 2 lactation stages. Data were split into multiple lactations stages for GWA, RHM, and transcriptomic analyses. ResultsStage-dependent GWAS and RHM identified 21 significant loci associated with DMI across multiple lactation stages. A total of 45 candidate genes were identified from GWA and RHM. Among all the 45 genes, six genes were later found significantly differently expressed between high and low feed intake animal groups using gene expression information from RNA-seq data. These genes show links to sugar and adipose metabolism, milk production, body weight, dopamine-reward pathways and immune functions. ConclusionsOur multi-omics analyses provide molecular evidence that the genetic basis of cattle DMI across lactation is not static. Temporal genomic variants associated with FI were identified with their transcriptomic patterns investigated, decoding the molecular mechanisms underlying DMI. Overall, the associated variants and candidate genes uncovered herein decoded genetic architecture of dairy feed intake on a temporal and multi-omics basis, enhancing the understanding of basic biology of dairy feed intake and informing breeding strategies aimed at improving dairy feed efficiency.

BackgroundBlack soldier fly larvae (Hermetia illucens, BSFL) efficiently bio-convert organic waste into high-value protein, which has significant potential in domesticated animal feed formulations. BSFL growth and bioconversion potential can be enhanced through selective breeding, which requires accurate estimates of genetic parameters and knowledge of genotype-by-diet (G x D) interactions. However, comprehensive knowledge of G x D interactions is limited, and reports of genetic parameters are sparse across genetic strains and production environments globally. ResultsThis study estimated heritabilities, dominance effects and genetic correlations for BSFL growth traits and quantified G x D interactions. Phenotypes of 2,097 fifth-instar larvae reared on three diets were recorded, including larval body weight (LBW), length (LL), width (LW), and surface area (LSA). All larvae were genotyped using a custom 6K Allegro SNP panel. Genetic parameters and G x D interactions were estimated by fitting an additive-dominance model in ASReml-R. Heritabilities for growth traits were low across diets (0.05-0.14), with diet-specific estimates ranging from low to moderate (0.06-0.36). Dominance effects were significant across the traits (0.09-0.19), and genetic correlations were high among growth traits (>0.81), except between LW and LL (0.51). G x D interactions were moderate across diets (-0.04-0.49). ConclusionResults suggest that moderate to high genetic gain is achievable over a long-term breeding programme, given the genetic basis of growth traits and BSFs short generation interval (38-45 days). However, G x D interactions must be considered, either through combined or diet-specific selection strategies, and the significant dominance effects suggest heterosis could accelerate improvement.

Cattle are broadly classified into two subspecies: Bos taurus, adapted to temperate climates, and Bos indicus, adapted to tropical environments. Indicine cattle show better heat tolerance and stronger disease resistance, whereas taurine cattle are known for higher milk yield and better meat quality. Improving milk yield while maintaining resistance to heat stress and infectious diseases is an important objective in cattle breeding. Marker-assisted selection is an effective approach for improving economically important traits, highlighting the need to understand genetic differences between taurine and indicine cattle. However, genome-wide comparisons between European taurine and Indian indicine cattle remain limited. To address this gap, whole-genome sequencing of 48 Indian indicine cattle was performed and combined with publicly available data. This enabled a comparative genomic analysis of 74 Indian indicine and 83 European taurine individuals to identify genes involved in heat tolerance and immune response. Genome-wide analyses using Fst and XP-CLR identified 4,343 and 1,457 differentiated genes, respectively, with 826 genes common to both methods. These genes were mainly associated with immune response, protein stability, and cytoskeletal structure. Strong selection signals were observed in three heat shock protein genes (DNAJC11, DNAJC5, and DNAJB11) and 229 immune-related genes. To examine the inheritance of these genes through crossbreeding, a haplotype-resolved genome assembly was generated for the Indian crossbreed Sunandini, which showed predominantly taurine ancestry (69.13-96.04%), with a smaller indicine contribution (1.22-1.87%). Several genes related to heat tolerance and immune response were inherited exclusively from indicine cattle, highlighting their importance for environmental adaptation and future breeding programs.

SummaryGhost introgression is a challenging problem in population genetics. Recent studies have explored supervised learning models, namely logistic regression and UNet++, to detect genomic footprints of ghost introgression. However, their applicability is limited because existing implementations are tailored to tasks in their respective publications, but not available as software implementations. Here, we present GAISHI, a Python package for identifying introgressed segments and alleles using machine learning and demonstrate its usage in a Human-Neanderthal introgression scenario. Availabity and implementationGAISHI is available on GitHub under the GNU General Public License v3.0. The source code can be found at https://github.com/xin-huang/gaishi.

Honeybee breeding plans are relatively recent in most countries. In France, diverse small-scale breeding groups are emerging. Beekeepers are highly diverse in their motivations, farm productions and services, practices and management techniques. Yet, little is known about what beekeepers would consider as relevant breeding goals in the design of breeding plans. We therefore conducted an online survey answered by about 250 French beekeepers, mostly professionals, to assess their perceived importance of including 20 pre-defined traits in breeding goals and to identify how beekeeping profiles might influence these priorities. Respondents rated each trait as essential, useful, or useless, and indicated if they wished useful or essential traits to be genetically improved or merely maintained at their current level. Results indicated a strong preference for multi-trait selection, with a median of 13 traits considered useful or essential. Honey yield, disease resistance, swarming tendency, gentleness, and summer feed autonomy, emerged as the main traits of interest with about 90% of beekeepers finding them at least useful. About 40% or more only wished to maintain these traits at their current level rather than to directionally improve them. A major exception to this was disease resistance, that 75% wanted to improve. Bees genetic background influenced the most the importance attributed to breeding goal traits, while other beekeeping profile characteristics only had a marginal effect on breeding goal trait priorities. Some poorly studied traits, such as summer and winter feed autonomy, winter diapause, and longevity, were considered at least useful in a breeding goal by over 70% of beekeepers. Future research is needed to explore possible selection criteria for these traits and estimate the potential for their genetic improvement. ImplicationsOur survey shows that French beekeepers wish to improve or maintain through selective breeding usual colony production and behavioral traits, but also colony resilience, especially disease resistance and feed autonomy. However, trait priorities differ depending on the genetic background of the bees used. This knowledge is essential for designing breeding programs that truly match beekeeper needs and for identifying which traits deserve research attention. In France, beekeepers are increasingly starting breeding efforts to adapt their bees to current conditions, facing growing pressures from climate change, diseases, invasive species, and pesticides. Well-designed breeding programs can support sustainable beekeeping and essential pollination services.

Current optimum contribution selection (OCS) implementations use point estimates of estimated breeding values (EBVs), potentially leading to suboptimal selections when individuals have uncertain genetic evaluations. We developed a framework assessing how EBV uncertainty affects OCS decisions through MCMC-based approaches using the COSMO optimizer in Julia, evaluated on Norway spruce (Picea abies, n=5,525) and Loblolly pine (Pinus taeda, n=926) populations. Agreement between point estimate (MAP-OCS) and MCMC-OCS was surprisingly low: mean overlap of only 26.6 (4.8) individuals in Norway spruce genotyped subpopulation and 14.1 (3.6) in full pedigree, with Loblolly pine intermediate at 16.0 (9.6). Despite this low individual-level agreement, selection frequency across MCMC iterations corresponded well with EBV rankings (Spearman{rho} = 0.782 for Norway spruce), confirming that higher-EBV individuals were preferentially selected under posterior uncertainty. To comprehensively quantify uncertainty impacts, we employed two complementary metrics: individual robustness scores measuring genetic gain stability upon candidate removal, and population-level contribution distribution metrics capturing concentration of genetic gain across selected individuals. Applying these metrics identified 25 high-risk individuals in Norway spruce and nine in Loblolly pine, and constrained exclusion of these individuals improved individual robustness by 16.5% in Loblolly pine (3.00% genetic gain loss) and 29.8% in Norway spruce (2.14% genetic gain loss). Our uncertainty-aware OCS framework successfully identifies unstable selections that may compromise long-term genetic gain, and we recommend assessing EBV uncertainty through posterior distributions and evaluating population-specific trade-offs when implementing uncertainty-aware selection strategies.