Genome-wide prediction (GWP) has become the state-of-the art method in artificial selection. Data sets often comprise number of genomic markers and individuals in ranges from a few thousands to millions. Hence, computational efficiency is important and various machine learning methods have successfully been used in GWP. Neural networks (NN) and deep learning (DL) are very flexible methods that usually show outstanding prediction properties on complex structured data, but their use in GWP is nevertheless rare and debated. This study describes a powerful NN method for genomic marker data that can easily be extended. It is shown that a one-dimensional convolutional neural network (CNN) can be used to incorporate the ordinal information between markers and, together with pooling and ℓ 1-norm regularization, provides a sparse and computationally efficient approach for GWP. The method, denoted CNNGWP, is implemented in the deep learning software Keras, and hyper-parameters of the NN are tuned with Bayesian optimization. Model averaged ensemble predictions further reduce prediction error. Evaluations show that CNNGWP improves prediction error by more than 25% on simulated data and around 3% on real pig data compared with results obtained with GBLUP and the LASSO. In conclusion, the CNNGWP provides a promising approach for GWP, but the magnitude of improvement depends on the genetic architecture and the heritability.
Pig herds in Africa comprise genotypes ranging from local ecotypes to commercial breeds. Many animals are composites of these two types and the best levels of crossbreeding for particular production systems are largely unknown. These pigs are managed without structured breeding programs and inbreeding is potentially limiting. The objective of this study was to quantify ancestry contributions and inbreeding levels in a population of smallholder pigs in Uganda. The study was set in the districts of Hoima and Kamuli in Uganda and involved 422 pigs. Pig hair samples were taken from adult and growing pigs in the framework of a longitudinal study investigating productivity and profitability of smallholder pig production. The samples were genotyped using the porcine GeneSeek Genomic Profiler (GGP) 50K SNP Chip. The SNP data was analyzed to infer breed ancestry and autozygosity of the Uganda pigs. The results showed that exotic breeds (modern European and old British) contributed an average of 22.8% with a range of 2–50% while “local” blood contributed 69.2% (36.9–95.2%) to the ancestry of the pigs. Runs of homozygosity (ROH) greater than 2 megabase (Mb) quantified the average genomic inbreeding coefficient of the pigs as 0.043. The scarcity of long ROH indicated low recent inbreeding. We conclude that the genomic background of the pig population in the study is a mix of old British and modern pig ancestries. Best levels of admixture for smallholder pigs are yet to be determined, by linking genotypes and phenotypic records.
The aim of this study was to conduct a multitrait 2-step approach applied to yield deviations and deregressed breeding values to get genetic parameters of functional longevity, clinical mastitis, early fertility disorders, cystic ovaries, and milk fever of Austrian Fleckvieh cattle. An approximate multitrait approach allows the combination of information from pseudo-phenotypes derived from different statistical models in routine genetic evaluation, which cannot be estimated easily in a full multitrait model. A total of 66,890 Fleckvieh cows were included in this study. For estimating genetic parameters, a simple linear animal model with year of birth as a fixed effect and animal as a random genetic effect was fitted. The joint analysis of yield deviations and deregressed breeding values was feasible. As expected, heritabilities were low, ranging from 0.03 (early fertility disorders) to 0.15 (functional longevity). Genetic correlations between functional longevity and clinical mastitis, early fertility disorders, cystic ovaries, and milk fever were 0.63, 0.29, 0.20, and 0.20, respectively. Within direct health traits genetic correlations were between 0.14 and 0.45. Results suggest that selecting for more robust disease-resistant cows would imply an improvement of functional longevity.
Welfare concerns for intensive beef production have often been raised, but on-farm welfare assessment studies are rare. The aim of this study was to apply the Welfare Quality® (WQ) welfare assessment system for fattening cattle on beef bull farms to evaluate the state of welfare at the level of WQ measures and of aggregated scores, as well as overall classification. In addition, the purpose was to evaluate two ways of providing feedback information to the farmers with regard to possible welfare improvements on the farms. The study was conducted in Austria, Germany and Italy on a total of 63 beef bull farms with deep litter or cubicle-housing systems. Assessments were carried out 3 times (1 month and 7 months apart from the initial visit). In every country, farmers were assigned to two treatment groups (feedback from initial visit as written report, F, written feedback plus oral advice, FA) and a control group (C), which did not receive any feedback. At the criterion level, the highest average welfare scores were obtained from ‘Absence of prolonged hunger’ (94/100 points) followed by ‘Absence of pain induced by management procedures’ (88/100) and ‘Comfort around resting’ (77/100). Most welfare concerns related to the criteria ‘Absence of disease’ (40/100), ‘Expression of social behaviour’ (44/100) and ‘Positive emotional state’ (48/100), thus indicating room for improvements. Two-thirds of the farms achieved the ‘Enhanced’ level, about one-third was judged ‘Acceptable’ and only one farm ‘Excellent’. After 6 months of monitoring period, there was no significant welfare improvement in both the treatment groups as compared with the control group. Reasons for the lack of effect may mainly be seen in the short monitoring period and a lack of external incentives. In conclusion, the WQ assessment system revealed areas for improvement, but longer term studies and investigations on alternative ways of transferring outcomes from on-farm welfare assessments to farmers should be carried out in future.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.