References of "Vandenplas, Jérémie"
     in
Bookmark and Share    
Full Text
See detailUse of high performance computing in animal breeding
Vandenplas, Jérémie ULg; Gengler, Nicolas ULg

Poster (2014, August 28)

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and ... [more ▼]

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and problem sizes. However, to our knowledge and especially for variance components estimations, most software available in animal breeding, based on sparse matrices computations, do not allow parallel computing and are limited by memory accessible by the central processing unit, or allow parallel computing only for options with dense matrices computations, which limits anyway problem sizes due to storage of dense matrices. The aim was to propose simple and effective modifications for the BLUPF90 family of programs to reduce computing time with consideration of required memory. Modifications were based on academic free packages proposing solver and sparse inversion for sparse symmetric indefinite linear systems. First, modifications concerned the sparse inversion subroutine implemented in the package FSPAK. Rearrangements of 'do' loops to allow optimizations of computer operations by some compilers and addition of OpenMP directives were performed. The ordering operation was modified to more easily compare a multiple minimum degree algorithm (MMD; implemented in FSPAK) and a multilevel nested dissection algorithm (implemented in METIS 4.0.3). Second, the package PARDISO Version 5.0.0 was used instead of FSPAK. This package proposes in particular a parallel solver and sparse inversion on shared-memory multiprocessors. Modified FSPAK and PARDISO were compared to original FSPAK using MMD through REMLF90. Different models, such as univariate or bivariate (random regressions) test-day animal and single-step genomic models, were tested. All jobs were run 5 times. With an appropriate ordering algorithm, speedup for each REMLF90 iteration were up to 7.5 for modified FSPAK and up to 22.8 for PARDISO with 2 threads. With 4 threads, speedup increased to 8.3 and 32.5, respectively. [less ▲]

Detailed reference viewed: 5 (0 ULg)
See detailBayesian approach integrating correlated foreign information into a multivariate genetic evaluation
Vandenplas, Jérémie ULg; Gengler, Nicolas ULg

in Book of Abstracts of the 65th Annual Meeting of the European Federation of Animal Science (2014, August 28)

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and ... [more ▼]

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and problem sizes. However, to our knowledge and especially for variance components estimations, most software available in animal breeding, based on sparse matrices computations, do not allow parallel computing and are limited by memory accessible by the central processing unit, or allow parallel computing only for options with dense matrices computations, which limits anyway problem sizes due to storage of dense matrices. The aim was to propose simple and effective modifications for the BLUPF90 family of programs to reduce computing time with consideration of required memory. Modifications were based on academic free packages proposing solver and sparse inversion for sparse symmetric indefinite linear systems. First, modifications concerned the sparse inversion subroutine implemented in the package FSPAK. Rearrangements of 'do' loops to allow optimizations of computer operations by some compilers and addition of OpenMP directives were performed. The ordering operation was modified to more easily compare a multiple minimum degree algorithm (MMD; implemented in FSPAK) and a multilevel nested dissection algorithm (implemented in METIS 4.0.3). Second, the package PARDISO Version 5.0.0 was used instead of FSPAK. This package proposes in particular a parallel solver and sparse inversion on shared-memory multiprocessors. Modified FSPAK and PARDISO were compared to original FSPAK using MMD through REMLF90. Different models, such as univariate or bivariate (random regressions) test-day animal and single-step genomic models, were tested. All jobs were run 5 times. With an appropriate ordering algorithm, speedup for each REMLF90 iteration were up to 7.5 for modified FSPAK and up to 22.8 for PARDISO with 2 threads. With 4 threads, speedup increased to 8.3 and 32.5, respectively. [less ▲]

Detailed reference viewed: 6 (3 ULg)
Peer Reviewed
See detailPrediction of Body Weight of Primiparous Dairy Cows Throughout Lactation
Vanrobays, Marie-Laure ULg; Vandenplas, Jérémie ULg; Hammami, Hedi ULg et al

in Proceedings of the 10th World Congress on Genetics Applied to Livestock Production (2014, August)

Body weight (BW) can be computed using linear conformation traits (CBW). However, these traits are recorded mostly once during a lactation. Therefore, predicted BW (PBW) is needed throughout the lactation ... [more ▼]

Body weight (BW) can be computed using linear conformation traits (CBW). However, these traits are recorded mostly once during a lactation. Therefore, predicted BW (PBW) is needed throughout the lactation (e.g., allowing feed intake prediction in milk recording systems). A two-step procedure was developed to obtain PBW using a random regression test-day model using CBW as observations. Added second step consisted in changing prior distribution for additive genetic random effects using results from first step to predict again PBW. This method was applied on 24,919 primiparous Holstein cows having 25,061 CBW to obtain PBW for 232,436 test-days. Results showed that applying both steps provided more accurate estimates than using only the first step. Furthermore, this procedure predicting PBW throughout lactation is also extremely flexible because actual BW can also be used together with CBW, the prediction model being able to accommodate different levels of accuracies. [less ▲]

Detailed reference viewed: 21 (9 ULg)
Full Text
See detailStrategies to combine novel traits across countries: example of heat stress
Hammami, Hedi ULg; Vandenplas, Jérémie ULg; Carabaño, Maria Jesus et al

Conference (2014, May 21)

Nowadays, novel traits are of great interest. However, phenotypes are siloed and mainly not shared. Heat stress is becoming problematic affecting animals’ performances and their well-being. Heat stress ... [more ▼]

Nowadays, novel traits are of great interest. However, phenotypes are siloed and mainly not shared. Heat stress is becoming problematic affecting animals’ performances and their well-being. Heat stress tolerance as a novel trait is only addressed by isolated within-country research studies. Integration and combination of local and foreign information sources is needed for better accuracy genetic evaluations. Therefore, this study was aimed to test the potential combination of sources of external information towards the evaluation of heat stress tolerance of dairy cattle. Long-term cow performances linked to environmental descriptors (weather parameters as proxy to climate change) collected over 10 years under the temperate conditions of the Walloon Region of Belgium and the hotter and warm Mediterranean conditions of Andalusia and Castile-La-Mancha Spanish regions were available. A total of 1,604,775 milk, fat, and protein test-day (TD) records linked to average daily temperature humidity (THI) values for 3-day lag before each TD were considered. Under a first strategy considering free-access to raw-data (phenotype and pedigree), a joint evaluation was firstly run using reaction norm models where production traits were considered as function of THI. A Belgian and a Spanish evaluation were also run using the same model. An alternative strategy considering only access to external information (i.e. regression coefficients for additive genetic effects (â and their associated REL)) was tested. In this case, foreign â and their REL resulting from the Spanish evaluation were first converted to the Belgian trait and thereafter integrated in the Belgian evaluation using a Bayesian approach. Rank correlations between regression coefficients, â (of the 1,104 bulls having daughters only in Spain) estimated by Belgian evaluation and â estimated by the joint evaluation were moderate (<=0.70). Corresponding rank correlations between â estimated by joint and Bayesian evaluations were significantly higher (ranging from 0.967 to 0.998), indicating that the Bayesian evaluation integrating external information was in good concordance with the joint evaluation. Results from this study indicated that the integration of external information via the Bayesian approach has a good potential to improve the genetic evaluation of sparse and siloed novel traits. [less ▲]

Detailed reference viewed: 20 (4 ULg)
Full Text
Peer Reviewed
See detailUnified method to integrate and blend several, potentially related, sources of information for genetic evaluation
Vandenplas, Jérémie ULg; Colinet, Frédéric ULg; Gengler, Nicolas ULg

in Genetics, Selection, Evolution (2014), 46

Background A condition to predict unbiased estimated breeding values by best linear unbiased prediction is to use simultaneously all available data. However, this condition is not often fully met. For ... [more ▼]

Background A condition to predict unbiased estimated breeding values by best linear unbiased prediction is to use simultaneously all available data. However, this condition is not often fully met. For example, in dairy cattle, internal (i.e. local) populations lead to evaluations based only on internal records while widely used foreign sires have been selected using internally unavailable external records. In such cases, internal genetic evaluations may be less accurate and biased. Because external records are unavailable, methods were developed to combine external information that summarizes these records, i.e. external estimated breeding values and associated reliabilities, with internal records to improve accuracy of internal genetic evaluations. Two issues of these methods concern double-counting of contributions due to relationships and due to records. These issues could be worse if external information came from several evaluations, at least partially based on the same records, and combined into a single internal evaluation. Based on a Bayesian approach, the aim of this research was to develop a unified method to integrate and blend simultaneously several sources of information into an internal genetic evaluation by avoiding double-counting of contributions due to relationships and due to records. Results This research resulted in equations that integrate and blend simultaneously several sources of information and avoid double-counting of contributions due to relationships and due to records. The performance of the developed equations was evaluated using simulated and real datasets. The results showed that the developed equations integrated and blended several sources of information well into a genetic evaluation. The developed equations also avoided double-counting of contributions due to relationships and due to records. Furthermore, because all available external sources of information were correctly propagated, relatives of external animals benefited from the integrated information and, therefore, more reliable estimated breeding values were obtained. Conclusions The proposed unified method integrated and blended several sources of information well into a genetic evaluation by avoiding double-counting of contributions due to relationships and due to records. The unified method can also be extended to other types of situations such as single-step genomic or multi-trait evaluations, combining information across different traits. [less ▲]

Detailed reference viewed: 29 (5 ULg)
Full Text
See detailBayesian approach integrating correlated foreign information into a multivariate genetic evaluation
Vandenplas, Jérémie ULg; Gengler, Nicolas ULg

Conference (2014)

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and ... [more ▼]

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and problem sizes. However, to our knowledge and especially for variance components estimations, most software available in animal breeding, based on sparse matrices computations, do not allow parallel computing and are limited by memory accessible by the central processing unit, or allow parallel computing only for options with dense matrices computations, which limits anyway problem sizes due to storage of dense matrices. The aim was to propose simple and effective modifications for the BLUPF90 family of programs to reduce computing time with consideration of required memory. Modifications were based on academic free packages proposing solver and sparse inversion for sparse symmetric indefinite linear systems. First, modifications concerned the sparse inversion subroutine implemented in the package FSPAK. Rearrangements of 'do' loops to allow optimizations of computer operations by some compilers and addition of OpenMP directives were performed. The ordering operation was modified to more easily compare a multiple minimum degree algorithm (MMD; implemented in FSPAK) and a multilevel nested dissection algorithm (implemented in METIS 4.0.3). Second, the package PARDISO Version 5.0.0 was used instead of FSPAK. This package proposes in particular a parallel solver and sparse inversion on shared-memory multiprocessors. Modified FSPAK and PARDISO were compared to original FSPAK using MMD through REMLF90. Different models, such as univariate or bivariate (random regressions) test-day animal and single-step genomic models, were tested. All jobs were run 5 times. With an appropriate ordering algorithm, speedup for each REMLF90 iteration were up to 7.5 for modified FSPAK and up to 22.8 for PARDISO with 2 threads. With 4 threads, speedup increased to 8.3 and 32.5, respectively. [less ▲]

Detailed reference viewed: 9 (0 ULg)
See detailUse of high performance computing in animal breeding
Vandenplas, Jérémie ULg; Gengler, Nicolas ULg

in Book of Abstracts of the 65th Annual Meeting of the European Federation of Animal Science (2014)

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and ... [more ▼]

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and problem sizes. However, to our knowledge and especially for variance components estimations, most software available in animal breeding, based on sparse matrices computations, do not allow parallel computing and are limited by memory accessible by the central processing unit, or allow parallel computing only for options with dense matrices computations, which limits anyway problem sizes due to storage of dense matrices. The aim was to propose simple and effective modifications for the BLUPF90 family of programs to reduce computing time with consideration of required memory. Modifications were based on academic free packages proposing solver and sparse inversion for sparse symmetric indefinite linear systems. First, modifications concerned the sparse inversion subroutine implemented in the package FSPAK. Rearrangements of 'do' loops to allow optimizations of computer operations by some compilers and addition of OpenMP directives were performed. The ordering operation was modified to more easily compare a multiple minimum degree algorithm (MMD; implemented in FSPAK) and a multilevel nested dissection algorithm (implemented in METIS 4.0.3). Second, the package PARDISO Version 5.0.0 was used instead of FSPAK. This package proposes in particular a parallel solver and sparse inversion on shared-memory multiprocessors. Modified FSPAK and PARDISO were compared to original FSPAK using MMD through REMLF90. Different models, such as univariate or bivariate (random regressions) test-day animal and single-step genomic models, were tested. All jobs were run 5 times. With an appropriate ordering algorithm, speedup for each REMLF90 iteration were up to 7.5 for modified FSPAK and up to 22.8 for PARDISO with 2 threads. With 4 threads, speedup increased to 8.3 and 32.5, respectively. [less ▲]

Detailed reference viewed: 8 (3 ULg)
See detailUse of high performance computing in animal breeding
Vandenplas, Jérémie ULg; Gengler, Nicolas ULg

in Book of Abstracts of the 65th Annual Meeting of the European Federation of Animal Science (2014)

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and ... [more ▼]

High-perfomance computing facilities proposing shared-memory and distributed-memory multiprocessors are becoming available. With those clusters, parallel computing could lead to increased performances and problem sizes. However, to our knowledge and especially for variance components estimations, most software available in animal breeding, based on sparse matrices computations, do not allow parallel computing and are limited by memory accessible by the central processing unit, or allow parallel computing only for options with dense matrices computations, which limits anyway problem sizes due to storage of dense matrices. The aim was to propose simple and effective modifications for the BLUPF90 family of programs to reduce computing time with consideration of required memory. Modifications were based on academic free packages proposing solver and sparse inversion for sparse symmetric indefinite linear systems. First, modifications concerned the sparse inversion subroutine implemented in the package FSPAK. Rearrangements of 'do' loops to allow optimizations of computer operations by some compilers and addition of OpenMP directives were performed. The ordering operation was modified to more easily compare a multiple minimum degree algorithm (MMD; implemented in FSPAK) and a multilevel nested dissection algorithm (implemented in METIS 4.0.3). Second, the package PARDISO Version 5.0.0 was used instead of FSPAK. This package proposes in particular a parallel solver and sparse inversion on shared-memory multiprocessors. Modified FSPAK and PARDISO were compared to original FSPAK using MMD through REMLF90. Different models, such as univariate or bivariate (random regressions) test-day animal and single-step genomic models, were tested. All jobs were run 5 times. With an appropriate ordering algorithm, speedup for each REMLF90 iteration were up to 7.5 for modified FSPAK and up to 22.8 for PARDISO with 2 threads. With 4 threads, speedup increased to 8.3 and 32.5, respectively. [less ▲]

Detailed reference viewed: 6 (2 ULg)
Full Text
Peer Reviewed
See detailShort communication: Alteration of priors for random effects in Gaussian linear mixed models
Vandenplas, Jérémie ULg; Christensen, Ole F.; Gengler, Nicolas ULg

in Journal of Dairy Science (2014), 97(9), 5880-5884

Linear mixed models, for which the prior multivariate normal distributions of random effects are assumed to have a mean equal to 0, are commonly used in animal breeding. However, some statistical analyses ... [more ▼]

Linear mixed models, for which the prior multivariate normal distributions of random effects are assumed to have a mean equal to 0, are commonly used in animal breeding. However, some statistical analyses (e.g., the consideration of a population under selection into a genomic scheme breeding, multiple-trait predictions of lactation yields, Bayesian approaches integrating external information into genetic evaluations) need to alter both the mean and (co)variance of the prior distributions and, to our knowledge, most software packages available in the animal breeding community do not permit such alterations. Therefore, the aim of this study was to propose a method to alter both the mean and (co)variance of the prior multivariate normal distributions of random effects of linear mixed models while using currently available software packages. The proposed method was tested on simulated examples with three different software packages available in animal breeding. The examples showed the possibility of the proposed method to alter both the mean and (co)variance of the prior distributions with currently available software packages through the use of an extended data file and a user supplied (co)variance matrix. [less ▲]

Detailed reference viewed: 15 (10 ULg)
Full Text
See detailStrategies to combine novel traits across countries: example of heat stress
Hammami, Hedi ULg; Vandenplas, Jérémie ULg; Carabaño, Maria Jesus et al

in Interbull Bulletin (2014), 48

Nowadays, novel traits are of great interest. However, phenotypes are siloed and mainly not shared. Heat stress is becoming problematic affecting animals’ performances and their well-being. Heat stress ... [more ▼]

Nowadays, novel traits are of great interest. However, phenotypes are siloed and mainly not shared. Heat stress is becoming problematic affecting animals’ performances and their well-being. Heat stress tolerance as a novel trait is only addressed by isolated within-country research studies. Integration and combination of local and foreign information sources is needed for better accuracy genetic evaluations. Therefore, this study was aimed to test the potential combination of sources of external information towards the evaluation of heat stress tolerance of dairy cattle. Long-term cow performances linked to environmental descriptors (weather parameters as proxy to climate change) collected over 10 years under the temperate conditions of the Walloon Region of Belgium and the hotter and warm Mediterranean conditions of Andalusia and Castile-La-Mancha Spanish regions were available. A total of 1,604,775 milk, fat, and protein test-day (TD) records linked to average daily temperature humidity (THI) values for 3-day lag before each TD were considered. Under a first strategy considering free-access to raw-data (phenotype and pedigree), a joint evaluation was firstly run using reaction norm models where production traits were considered as function of THI. A Belgian and a Spanish evaluation were also run using the same model. An alternative strategy considering only access to external information (i.e. regression coefficients for additive genetic effects (â and their associated REL)) was tested. In this case, foreign â and their REL resulting from the Spanish evaluation were first converted to the Belgian trait and thereafter integrated in the Belgian evaluation using a Bayesian approach. Rank correlations between regression coefficients, â (of the 1,104 bulls having daughters only in Spain) estimated by Belgian evaluation and â estimated by the joint evaluation were moderate (<=0.70). Corresponding rank correlations between â estimated by joint and Bayesian evaluations were significantly higher (ranging from 0.967 to 0.998), indicating that the Bayesian evaluation integrating external information was in good concordance with the joint evaluation. Results from this study indicated that the integration of external information via the Bayesian approach has a good potential to improve the genetic evaluation of sparse and siloed novel traits. [less ▲]

Detailed reference viewed: 35 (1 ULg)
Full Text
See detailCombinaison d'informations génomiques, phénotypiques et généalogiques
Vandenplas, Jérémie ULg; Gengler, Nicolas ULg

Conference given outside the academic context (2013)

Detailed reference viewed: 15 (7 ULg)
Full Text
Peer Reviewed
See detailGenetic variance in micro-environmental sensitivity for milk and milk quality in Walloon Holstein cattle
Vandenplas, Jérémie ULg; Bastin, Catherine ULg; Gengler, Nicolas ULg et al

in Journal of Dairy Science (2013), 96

Animals that are robust to environmental changes are desirable in the current dairy industry. Genetic differences in micro-environmental sensitivity can be studied through heterogeneity of residual ... [more ▼]

Animals that are robust to environmental changes are desirable in the current dairy industry. Genetic differences in micro-environmental sensitivity can be studied through heterogeneity of residual variance between animals. However, residual variance between animals is usually assumed homogeneous in traditional genetic evaluations. The aim of this study was to investigate genetic heterogeneity of residual variance by estimating variance components in residual variance for milk yield, somatic cell score, contents in milk (g/dL) of two groups of milk fatty acids (i.e. saturated fatty acids and unsaturated fatty acids) and the content in milk of one individual fatty acid (i.e. the oleic acid, C18:1 cis-9), for first-parity Holstein cows in the Walloon Region of Belgium. A total of 146,027 test-day records from 26,887 cows in 747 herds were available. All cows had at least three records and had a known sire. These sires had at least 10 cows with records and each herd x test-day had at least five cows. The five traits were analyzed separately based on fixed lactation curve and random regression test-day models for the mean. Estimation of variance components was performed by running iteratively Expectation Maximization-Restricted Maximum Likelihood algorithm by the implementation of double hierarchical generalized linear models. Based on fixed lactation curve test-day mean models, heritability for residual variances ranged between 1.01*10-3 and 4.17*10-3 for all traits. The genetic standard deviation in residual variance (i.e. approximately the genetic coefficient of variation of residual variance) ranged between 0.12 and 0.17. Therefore, some genetic variance in micro-environmental sensitivity existed in the Walloon Holstein dairy cattle for the five studied traits. The standard deviations due to herd x test-day and permanent environment in residual variance ranged between 0.36 and 0.45 for herd x test-day effect and between 0.55 and 0.97 for permanent environmental effect. Therefore, non-genetic effects also contributed substantially to the micro-environmental sensitivity. Results also showed that the addition of random regressions to the mean model did not reduce heterogeneity in residual variance and that genetic heterogeneity of residual variance was not simply an effect of an incomplete mean model. [less ▲]

Detailed reference viewed: 72 (48 ULg)
Full Text
See detailGenetics of mastitis in the Walloon Region of Belgium
Bastin, Catherine ULg; Vandenplas, Jérémie ULg; Laine, Aurélie ULg et al

in Book of Abstracts of the 64th Annual Meeting of the European Federation of Animal Science (2013, August)

Detailed reference viewed: 31 (19 ULg)
Full Text
See detailGenetics of mastitis in the Walloon Region of Belgium
Bastin, Catherine ULg; Vandenplas, Jérémie ULg; Laine, Aurélie ULg et al

Conference (2013, August)

Detailed reference viewed: 24 (8 ULg)
Full Text
See detailDevelopment of a genomic evaluation for milk production for a local bovine breed
Colinet, Frédéric ULg; Vandenplas, Jérémie ULg; Faux, Pierre ULg et al

in Book of Abstracts of the 64th Annual Meeting of the European Federation of Animal Science (2013, August)

Detailed reference viewed: 24 (10 ULg)