Leveraging Breeding Values Obtained from Random Regression Models for Genetic Inference of Longitudinal Traits

Abstract The objective of this work was to verify the gain in reliability of estimated breeding values (EBVs), when random regression models are applied instead of conventional 305-day lactation models, using fat and protein yield records of Brazilian Holstein cattle for future genetic evaluations. Data set contained 262,426 test-day fat and protein yield records, and 30,228 fat and protein lactation records at 305 days from first lactation. Single trait random regression models using Legendre polynomials and single trait lactation models were applied. Heritability for 305-day yield from lactation models was 0.24 (fat) and 0.17 (protein), and from random regression models was 0.20 (fat) and 0.21 (protein). Spearman correlations of EBVs, between lactation models and random regression models, for 305-day yield, ranged from 0.86 to 0.97 and 0.86 to 0.98 (bulls), and from 0.80 to 0.89 and 0.81 to 0.86 (cows), for fat and protein, respectively. Average increase in reliability of EBVs for 305-day yield of bulls ranged from 2 to 16% (fat) and from 4 to 26% (protein), and average reliability of cows ranged from 24 to 38% (fat and protein), which is higher than in the lactation models. Random regression models using Legendre polynomials will improve genetic evaluations of Brazilian Holstein cattle due to the reliability increase of EBVs, in comparison with 305-day lactation models.

Download Full-text

Development of genetic models for estimation of racing performances in German thoroughbreds

Archives Animal Breeding ◽

10.5194/aab-47-505-2004 ◽

2004 ◽

Vol 47 (6) ◽

pp. 505-516

Author(s):

A.-E. Bugislaus ◽

R. Roehe ◽

H. Uphaus ◽

E. Kalm

Keyword(s):

Animal Model ◽

Linear Regression ◽

Statistical Models ◽

Regression Models ◽

Genetic Variance ◽

Fixed Effect ◽

Random Regression ◽

Square Root ◽

Breeding Values ◽

High Performing

Abstract. The objective of this study was to develop new statistical models for genetic estimation of racing performances in German thoroughbreds. Analysed performance traits were "square root of rank at finish", "square root of distance to first placed horse in a race" and "log of earnings". These traits were found to be influenced by the carried weight, which was determined by the horse's earlier performance. Therefore, new traits were developed based on random regression models, which were independent from the carried weights. Heritabilities were first estimated for these created traits "new rank at finish" (h2 = 0.101) and "new distance to first placed horse in a race" (h2 = 0.142) by using two univariate animal models. When considering a linear regression of carried weights as fixed effect in the statistical model, heritabilities for "square root of rank at finish" (h2 = 0.086) and "square root of distance to first placed horse in a race" (h2 = 0.124) decreased. Breeding values of “new rank at finish” and "new distance to first placed horse in a race" were compared with breeding values of "square root of rank at finish" and "square root of distance to first placed horse in a race", in which carried weight was considered as fixed regression in the model. These two different models were compared by two criteria. Breeding values were overestimated for low performing thoroughbreds and underestimated for high performing horses when considering a linear regression of carried weights as fixed effect in the model. Statistical models considering new created traits ("new rank at finish" and "new distance to first placed horse in a race") which were independent of carried weights, showed better suitability for genetic estimation. Due to high genetic correlation with other traits and showing highest genetic variance a univariate animal model for the trait “new distance to first placed horse in a race” was recommended for genetic estimation.

Download Full-text

PSXV-1 Genetic evaluation of longevity of cows culled due to fertility-related problems using random regression models and censored data

Journal of Animal Science ◽

10.1093/jas/skab235.477 ◽

2021 ◽

Vol 99 (Supplement_3) ◽

pp. 261-261

Author(s):

Hinayah R Oliveira ◽

Stephen P Miller ◽

Luiz F Brito ◽

Flavio S Schenkel

Keyword(s):

Censored Data ◽

Regression Models ◽

Genetic Evaluation ◽

Random Regression ◽

Breeding Values ◽

Structural Problems ◽

Time Period ◽

And Performance ◽

Genetic Evaluations ◽

The Impact

Abstract A recent study showed that longevity based on different culling reasons should be considered as different traits in genetic evaluations. However, it is still necessary to create a pipeline that avoid including/excluding animals culled for different reasons in every genetic evaluation run. This study aimed to: 1) perform a genetic evaluation of the longevity of cows culled due to fertility-related problems including records of animals culled for other reasons (i.e., age, structural problems, disease, and performance) as censored records; and, 2) identify the impact of censored data in the genetic parameters and breeding values estimated. Two longevity indicators were evaluated: traditional (TL; time from first calving to culling) and functional (FL; time period in which the cow was alive and also calving after its first calving) longevity. Both TL and FL were evaluated from 2 to 15 years-old, and codified as binary traits for each age (0 = culled and 1 = alive/calved). Both trait definitions were analyzed using a Bayesian random regression linear model. Animals culled for reasons other than fertility were either excluded from the data (standard) or had their records censored after the culling date reported in the dataset (censored). After the quality control, 154,419 and 450,124 animals had uncensored and censored records, respectively. Heritabilities estimated for TL over the ages ranged from 0.02 to 0.13 for standard, and from 0.01 to 0.12 for censored datasets. Heritabilities estimated for FL ranged from 0.01 to 0.14 (standard), and from 0.01 to 0.13 (censored). Average (SD) correlation of breeding values predicted over all ages, using the standard and censored datasets, was 0.77 (0.16) for TL, and 0.83 (0.11) for FL. Our findings suggest that including censored data in the analyses might impact the genomic evaluations and further work is need to determine the optimal predictive approach.

Download Full-text

Predicting breeding values for milk yield of Guzerá (Bos indicus) cows using random regression models

Livestock Science ◽

10.1016/j.livsci.2014.05.023 ◽

2014 ◽

Vol 167 ◽

pp. 41-50 ◽

Cited By ~ 4

Author(s):

D.J.A. Santos ◽

M.G.C.D. Peixoto ◽

R.R. Aspilcueta Borquis ◽

J.C.C. Panetto ◽

L. El Faro ◽

...

Keyword(s):

Milk Yield ◽

Regression Models ◽

Bos Indicus ◽

Random Regression ◽

Breeding Values

Download Full-text

Analysis of censored survival data using random regression models

Animal Science ◽

10.1017/s1357729800055491 ◽

2001 ◽

Vol 72 (1) ◽

pp. 1-10 ◽

Cited By ~ 40

Author(s):

R. F. Veerkamp ◽

S. Brotherstone ◽

B. Engel ◽

T. H. E. Meuwissen

Keyword(s):

Milk Yield ◽

Survival Data ◽

Variance Components ◽

Regression Models ◽

Proportional Hazards ◽

Random Regression ◽

Breeding Values ◽

Censored Survival Data ◽

Proportional Hazards Models ◽

Randomly Censored Data

AbstractCensoring of records is a problem in the prediction of breeding values for longevity, because breeding values are required before actual lifespan is known. In this study we investigated the use of random regression models to analyse survival data, because this method combines some of the advantages of a multitrait approach and the more sophisticated proportional hazards models. A model was derived for the binary representation of survival data and links with proportional hazards models and generalized linear models are shown. Variance components and breeding values were predicted using a linear approximation, including time-dependent fixed effects and random regression coefficients. Production records in lactations 1 to 5 were available on 24741 cows in the UK, all having had the opportunity to survive five lactations. The random regression model contained a linear regression on milk yield within herd (no. = 1417) by lactation number (no. = 4), Holstein percentage and year-month of calving effect (no. = 72). The additive animal genetic effects were modelled using orthogonal polynomials of order 1 to 4 with random coefficients and the error terms were fitted for each lactation separately, either correlated or not. Variance components from the full (i.e. uncensored) data set, were used to predict breeding values for survival in each lactation from both uncensored and randomly censored data. In the uncensored data, estimates of heritabilities for culling probability in each lactation ranged from 0·02 to 0·04. Breeding values for lifespan (calculated from the survival breeding values) had a range of 2·4 to 3·6 lactations and a standard deviation of 0·25. Correlations between predicted breeding values for 129 bulls, each with more than 30 daughters, from the various data sets ranged from 0·81 to 0·99 and were insensitive to the model used. It is concluded that random regression analysis models used for test-day records analysis of milk yield, might also be of use in the analysis of censored survival data.

Download Full-text

High-Throughput Phenotyping and Random Regression Models Reveal Temporal Genetic Control of Soybean Biomass Production

Frontiers in Plant Science ◽

10.3389/fpls.2021.715983 ◽

2021 ◽

Vol 12 ◽

Author(s):

Fabiana Freitas Moreira ◽

Hinayah Rojas de Oliveira ◽

Miguel Angel Lopez ◽

Bilal Jamal Abughali ◽

Guilherme Gomes ◽

...

Keyword(s):

High Throughput ◽

Regression Models ◽

Genetic Architecture ◽

Crop Improvement ◽

Random Regression ◽

Genome Wide Association Studies ◽

Breeding Values ◽

High Throughput Phenotyping ◽

General Response ◽

Over Time

Understanding temporal accumulation of soybean above-ground biomass (AGB) has the potential to contribute to yield gains and the development of stress-resilient cultivars. Our main objectives were to develop a high-throughput phenotyping method to predict soybean AGB over time and to reveal its temporal quantitative genomic properties. A subset of the SoyNAM population (n = 383) was grown in multi-environment trials and destructive AGB measurements were collected along with multispectral and RGB imaging from 27 to 83 days after planting (DAP). We used machine-learning methods for phenotypic prediction of AGB, genomic prediction of breeding values, and genome-wide association studies (GWAS) based on random regression models (RRM). RRM enable the study of changes in genetic variability over time and further allow selection of individuals when aiming to alter the general response shapes over time. AGB phenotypic predictions were high (R2 = 0.92–0.94). Narrow-sense heritabilities estimated over time ranged from low to moderate (from 0.02 at 44 DAP to 0.28 at 33 DAP). AGB from adjacent DAP had highest genetic correlations compared to those DAP further apart. We observed high accuracies and low biases of prediction indicating that genomic breeding values for AGB can be predicted over specific time intervals. Genomic regions associated with AGB varied with time, and no genetic markers were significant in all time points evaluated. Thus, RRM seem a powerful tool for modeling the temporal genetic architecture of soybean AGB and can provide useful information for crop improvement. This study provides a basis for future studies to combine phenotyping and genomic analyses to understand the genetic architecture of complex longitudinal traits in plants.

Download Full-text

Modeling of average growth curve in Santa Ines sheep using random regression models

Revista Brasileira de Zootecnia ◽

10.1590/s1516-35982011000200012 ◽

2011 ◽

Vol 40 (2) ◽

pp. 314-322 ◽

Cited By ~ 4

Author(s):

José Lindenberg Rocha Sarmento ◽

Robledo de Almeida Torres ◽

Wandrick Hauss de Sousa ◽

Lucia Galvão de Albuquerque ◽

Raimundo Nonato Braga Lôbo ◽

...

Keyword(s):

Growth Curve ◽

Regression Models ◽

Random Regression ◽

Average Growth ◽

Breeding Values ◽

Cubic Polynomial ◽

Linear Fit ◽

Santa Inês ◽

The Difference ◽

Cubic Function

Polynomial functions of age of different orders were evaluated in the modeling of the average growth trajectory in Santa Ines sheep in random regression models. Initially, the analyses were performed not considering the animal effect. Subsequently, the random regression analyses were performed including the random effects of the animal and its mother (genetic and permanent environment). The linear fit was lower, and the other orders were similar until near 100 days of age. The cubic function provided the closest fit of the observed averages, mainly at the end of the curve. Orders superior to this one tended to present incoherent behavior with the observed weights. The estimated direct heritabilities, considering the linear fit, were higher to those estimated by considering other functions. The changes in animal ranking based on predicted breeding values using linear fit and superior orders were small; however, the difference in magnitude of the predicted breeding values was higher, reaching values 77% higher than those obtained with the cubic function. The cubic polynomial function is efficient in describing the average growth curve.

Download Full-text

Comparison of Breeding Values for Daily Gains of Bulls Estimated with Multi-Trait and Random Regression Models

Archives Animal Breeding ◽

10.5194/aab-50-147-2007 ◽

2007 ◽

Vol 50 (2) ◽

pp. 147-154

Author(s):

H. Krejčová ◽

N. Mielenz ◽

J. Přibyl ◽

L. Schüler

Keyword(s):

Legendre Polynomials ◽

Regression Models ◽

Reference Model ◽

Average Daily Gain ◽

Regression Coefficients ◽

Random Regression ◽

Breeding Values ◽

Higher Rank ◽

Estimated Breeding Values ◽

Daily Gain

Abstract. In this study, random regression models with Legendre polynomials of the 2nd, 3rd and 4th degree (RR2, RR3 and RR4) are compared with regard to the estimation of breeding values for the average daily gain of Czech Pied bulls (Simmental type). The data were prepared such that a multi-trait model (MTM) could be used as reference model. For each bull, 8 repeated records or fewer were available for the testing period from the 12th to the 420th day of life. For the modeling of the expected value structure, the fixed regression coefficients of the Legendre polynomials were subordinated hierarchically to the herd-year-season effects (HYS). For the comparison of the random regression models with the reference model, rank correlations between the estimated breeding values of various animal groups were calculated and a variety of top-lists were analyzed. In general, models RR3 and RR4 returned higher rank correlations with MTM in comparison to model RR2. Additionally, the number of common animals in the 1% and 10% top-lists showed that models RR3 and RR4 are to be preferred over RR2 when it comes to the estimation of breeding values.

Download Full-text

Leveraging breeding values obtained from random regression models for genetic inference of longitudinal traits

10.1101/435685 ◽

2018 ◽

Author(s):

Malachy T Campbell ◽

Mehdi Momen ◽

Harkamal Walia ◽

Gota Morota

Keyword(s):

Genetic Analysis ◽

Genomic Prediction ◽

Shoot Growth ◽

Effective Means ◽

Repeated Measurements ◽

Random Regression ◽

Growth Trajectories ◽

Vegetative Development ◽

Breeding Values ◽

Genetic Inference

Understanding the genetic basis of dynamic plant phenotypes has largely been limited due to lack of space and labor resources needed to record dynamic traits, often destructively, for a large number of genotypes. However, the recent advent of image-based phenotyping platforms has provided the plant science community with an effective means to non-destructively evaluate morphological, developmental, and physiological processes at regular, frequent intervals for a large number of plants throughout development. The statistical frameworks typically used for genetic analyses (e.g. genome-wide association mapping, linkage mapping, and genomic prediction) in plant breeding and genetics are not particularly amenable for repeated measurements. Random regression (RR) models are routinely used in animal breeding for the genetic analysis of longitudinal traits, and provide a robust framework for modeling traits trajectories and performing genetic analysis simultaneously. We recently used a RR approach for genomic prediction of shoot growth trajectories in rice using 33,674 SNPs. In this study, we have extended this approach for genetic inference by leveraging genomic breeding values derived from RR models for rice shoot growth during early vegetative development. This approach provides improvements over a conventional single time point analyses for discovering loci associated with shoot growth trajectories. The RR approach uncovers persistent, as well as time-specific, transient quantitative trait loci. This methodology can be widely applied to understand the genetic architecture of other complex polygenic traits with repeated measurements.

Download Full-text