Systematic Evaluation of Genomic Prediction Algorithms for Genomic Prediction and Breeding of Aquatic Animals

Genes, Journal Year: 2022, Volume and Issue: 13(12), P. 2247 - 2247

Published: Nov. 29, 2022

The extensive use of genomic selection (GS) in livestock and crops has led to a series genomic-prediction (GP) algorithms despite the lack single algorithm that can suit all species traits. A systematic evaluation available GP is thus necessary identify optimal for selective breeding aquaculture species. In this study, comparison ten algorithms, including both traditional machine-learning was conducted using publicly genotype phenotype data eight traits, weight disease resistance from five study aimed provide insights into aquatic animals. Notably, no showed best performance However, reproducing kernel Hilbert space (RKHS) support-vector machine (SVM) achieved relatively high prediction accuracies most tested Bayes random forest (RF) better prevented noise interference phenotypic compared other algorithms. performances Crassostrea gigas dataset were improved by genome-wide association (GWAS) select subsets significant SNPs. An R package, “ASGS,” which integrates commonly used efficiently finding algorithm, developed assist application This work provides valuable information tool optimizing GP, aiding genetic

Language: Английский

---

Artificial intelligence in aquaculture: basis, applications, and future perspectives

Scientia Agropecuaria, Journal Year: 2022, Volume and Issue: 13(1), P. 79 - 96

Published: March 28, 2022

Advances in data management technologies are being adapted to resolve difficulties and impacts that aquaculture manifests, some aspects over the years have not been fully managed, now more feasible solve, such as optimization of variables intervene growth increase biomass, prediction water quality parameters manage make decisions during farming fish, evaluation environment impact generated by aquaculture, diagnosis diseases fish determine specific treatments, handling, closure farms. The objective this article was review within last 20 various techniques, methodologies, models, algorithms, software, devices used artificial intelligence, machine learning deep systems, solve a simpler way, quickly precisely manifests. In addition, fundamentals automatic explained, well recommendations for future study on areas interest reduction production costs through feeding based good practices quality, identification sex do present sexual dimorphism, determination attributes degree pigmentation salmon trout.

Language: Английский

---

Machine Learning Based Approaches for Livestock Symptoms and Diseases Prediction and Classification

Published: May 9, 2024

Language: Английский

---

Multivariate genomic prediction for commercial traits of economic importance in Banana shrimp Fenneropenaeus merguiensis

Aquaculture, Journal Year: 2022, Volume and Issue: 555, P. 738229 - 738229

Published: April 9, 2022

Language: Английский

---

Accuracies of genomic predictions for disease resistance of striped catfish to Edwardsiella ictaluri using artificial intelligence algorithms

G3 Genes Genomes Genetics, Journal Year: 2021, Volume and Issue: 12(1)

Published: Oct. 22, 2021

Assessments of genomic prediction accuracies using artificial intelligent (AI) algorithms (i.e., machine and deep learning methods) are currently not available or very limited in aquaculture species. The principal aim this study was to examine the predictive performance these new methods for disease resistance Edwardsiella ictaluri a population striped catfish Pangasianodon hypophthalmus make comparisons with four common methods, i.e., pedigree-based best linear unbiased (PBLUP), genomic-based (GBLUP), single-step GBLUP (ssGBLUP) nonlinear Bayesian approach (notably BayesR). Our analyses ML-KAML) DL-MLP DL-CNN) together (PBLUP, GBLUP, ssGBLUP, BayesR) were conducted two main traits survival status coded as 0 1 time, days that animals still alive after challenge test) pedigree consisting 560 individual (490 offspring 70 parents) genotyped 14,154 single nucleotide polymorphism (SNPs). results 6,470 SNPs quality control showed outperformed PBLUP, increases both by 9.1-15.4%. However, obtained from comparable those estimated BayesR. Imputation missing genotypes AlphaFamImpute increased 5.3-19.2% all data used. On other hand, there insignificant decreases (0.3-5.6%) time when multivariate models used comparison univariate analyses. Interestingly, based on only highly significant (P < 0.00001, 318-400 1,362-1,589 time) somewhat lower (0.3-15.6%) than whole set SNPs. In most our analyses, higher (0/1 data). It is concluded although prospects application selection increase E. breeding programs, further evaluation should be made independent families/populations more accumulated future generations avoid possible biases genetic parameters estimates disease-resistant studied P. hypophthalmus.

Language: Английский

---

Viral nervous necrosis resistance in gilthead sea bream (Sparus aurata) at the larval stage: heritability and accuracy of genomic prediction with different training and testing settings

Genetics Selection Evolution, Journal Year: 2023, Volume and Issue: 55(1)

Published: April 3, 2023

The gilthead sea bream (Sparus aurata) has long been considered resistant to viral nervous necrosis (VNN), until recently, when significant mortalities caused by a reassortant virus (NNV) strain were reported. Selective breeding enhance resistance against NNV might be preventive action. In this study, 972 larvae subjected challenge test and the symptomatology was recorded. All experimental fish their parents genotyped using genome-wide single nucleotide polymorphism (SNP) array consisting of over 26,000 markers.Estimates pedigree-based genomic heritabilities VNN consistent with each other (0.21, highest posterior density interval at 95% (HPD95%): 0.1-0.4; 0.19, HPD95%: 0.1-0.3, respectively). association study suggested one region, i.e., in linkage group (LG) 23 that involved resistance, although it far from significance threshold. accuracies (r) predicted estimated values (EBV) provided three Bayesian regression models (Bayes B, Bayes C, Ridge Regression) on average equal 0.90 assessed set cross-validation (CV) procedures. When relationships between training testing sets minimized, accuracy decreased greatly (r = 0.53 for validation based clustering, r 0.12 leave-one-family-out approach focused challenged fish). Classification phenotype predictions or pedigree-based, all data included, EBV as classifiers moderately accurate (area under ROC curve 0.60 0.66, respectively).The estimate heritability indicates is feasible implement selective programs increased larvae/juveniles. Exploiting information offers opportunity developing prediction tools can trained phenotypes, minimal differences classification performance trait phenotype. long-term view, weakening ties animals leads accuracies, thus periodical update reference population new mandatory.

Language: Английский

---

Using Bayesian threshold model and machine learning method to improve the accuracy of genomic prediction for ordered categorical traits in fish

Agriculture Communications, Journal Year: 2023, Volume and Issue: 1(1), P. 100005 - 100005

Published: July 4, 2023

Ordered categorical traits are commonly used in fish breeding programs as they easier to obtain than continuous observations. However, most studies treat ordered linear and analyze them using models, which can lead a serious reduction prediction accuracy by violating the basic assumptions of models. The aim this study was evaluate advantages Bayesian threshold model machine learning method genomic fish. based on analyses simulated data real Atlantic salmon. Ordinal were with varying numbers categories (2, 3 4) levels heritabilities (0.1, 0.3 0.5). Linear models BayesA BayesCπ methods, well method, support vector regression default (SVRdef) tuning (SVRtuning) hyperparameters investigate their abilities. results showed that yielded 2.1%, 2.6% 2.9% higher accuracies average for 2-, 3- 4-category traits, respectively, Furthermore, SVRtuning produced compared SVRdef all scenarios. For data, 1.2% 3.3% 6.6% respectively. In conclusion, use beneficial

Language: Английский

---

Evaluation of machine learning method in genomic selection for growth traits of Pacific white shrimp

Aquaculture, Journal Year: 2023, Volume and Issue: 581, P. 740376 - 740376

Published: Nov. 15, 2023

Language: Английский

---

An Integrated GIS-Based Reinforcement Learning Approach for Efficient Prediction of Disease Transmission in Aquaculture

Information, Journal Year: 2023, Volume and Issue: 14(11), P. 583 - 583

Published: Oct. 24, 2023

This study explores the design and capabilities of a Geographic Information System (GIS) incorporated with an expert knowledge system, tailored for tracking monitoring spread dangerous diseases across collection fish farms. Specifically targeting aquacultural regions Greece, system captures geographical climatic data pertinent to these A feature this is its ability calculate disease transmission intervals between individual cages broader farm entities, providing crucial insights into dynamics. These then act as entry point our system. To enhance predictive precision, we employed various machine learning strategies, ultimately focusing on reinforcement (RL) environment. RL framework, enhanced by Multi-Armed Bandit (MAB) technique, stands out powerful mechanism effectively managing flow virus transmissions within Empirical tests highlight efficiency MAB approach, which, in direct comparisons, consistently outperformed other algorithmic options, achieving impressive accuracy rate 96%. Looking ahead future work, plan integrate buffer techniques delve deeper advanced models current The results set stage research modeling aquaculture health management, aim extend even further.

Language: Английский

---

Smart Low-Cost Control System for Fish Farm Facilities

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(14), P. 6244 - 6244

Published: July 18, 2024

Projections indicate aquaculture will produce 106 million tonnes of fish by 2030, emphasizing the need for efficient and sustainable practices. New technologies can provide a valuable tool adequate farm management. The aim this paper is to explore factors affecting well-being, design control systems aquaculture, proposal smart system based on algorithms improve efficiency sustainability. First, we identify domains well-being: production domain, abiotic biotic domain. Then, evaluate interactions between elements present in each domain key aspects be monitored. This conducted two types farming facilities: cages sea recirculating systems. A total 86 have been identified, which 17 32 were selected included monitoring series are proposed optimize We predefined algorithms, energy-efficient fault tolerance data management algorithm. considering all aforementioned factors, scenarios simulated benefits In case, turbidity when algorithm used represents 12.5% not used. Their use resulted 35% reduction energy consumption aerator was implemented.

Language: Английский

---