Search

International audience; Breeding for resilience requires a better understanding of intra-flock variability and the related mechanisms responsible for robustness traits. Among such traits, the animals’ ability to cope with feed fluctuations by mobilizing or restoring body reserves (BR) is a key mechanism in ruminants. The objective of this work was to characterize individual variability in BR dynamics in productive Romane ewes reared in extensive conditions. The BR dynamics profiles were characterized by combining individual longitudinal measurements of BW and body condition scores (BCS) over several production cycles. Historical data, including up to 2628 records per trait distributed in 1146 ewes, underwent cluster analysis. Two to four trajectories were observed for BW depending on the cycle, while three trajectories were found for BCS, whatever the cycle. Most trajectories suggested that BR dynamics were similar but the level of BR may differ between ewes. Nevertheless, some trajectories suggested that both BR dynamics and levels were different for a proportion of ewes. Clustering on BW and BCS profiles adjusted for individual level trends, resulted in differences only in the level of BW or BCS, rather than differences in trajectories. Thus, the overall shape of trajectories was not changed considering or not the individual level trend across cycles. In addition to individual variability, the ewe’s age at first lambing and litter size contributed to the distribution of the ewes between the trajectories. Regarding the entire productive life, three trajectories were observed for BW and BCS changes over three productive cycles. Increase in BW at each cycle suggested that ewes kept growing up until 3 to 4 years old in our conditions. Similar alternation of BCS gains and losses across cycles suggested BR dynamics might be repeatable. Many individual trajectories remained the same throughout a ewe’s life, whatever the age at first lambing, parity or litter size. Our results demonstrate the relevance of using BW and BCS changes for characterizing the diversity of BR mobilization–accretion profiles in sheep in a long timespan perspective.

International audience; A high level of production at the peak of lactation may be associated with animal health disorders, high feeding costs, and reduced milk supply throughout the year. The objective of this study was to typologize the lactation curves in French dairy goats and analyze the influence of environmental and genetic factors on these curves. The data set consisted of 2,231,720 monthly test-day records of 213,534 French Saanen and Alpine goats recorded between September 2008 and June 2012. First, principal component analysis classified the shape of the lactation curves into 3 principal components: the first component accounted for milk yield level throughout lactation, the second component accounted for lactation persistency, and the third component accounted for milk yield in mid-lactation. Then, from the principal component scores, the lactations were clustered into 5 different groups. Most lactations had a similar shape to the mean curve, except 30% of the lactations that fell into 3 clusters that had a high production level at the peak and then a different persistency according to cluster. Estimated breeding value for milk yield and home region of breeding were the factors most related to lactation production level. Month of kidding, breed, and gestation stage had the biggest effect on persistency. Month of kidding was the factor most strongly linked to mid-lactation production. A herd effect was observed on all 3 principal components.

Background Goats were domesticated 10,500 years ago to supply humans with useful resources. Since then, specialized breeds that are adapted to their local environment have been developed and display specific genetic profiles. The VarGoats project is a 1000 genomes resequencing program designed to cover the genetic diversity of the Capra genus. In this study, our main objective was to assess the use of sequence data to detect genomic regions associated with traits of interest in French Alpine and Saanen breeds. Results Direct imputation from the GoatSNP50 BeadChip genotypes to sequence level was investigated in these breeds using FImpute and different reference panels: within-breed, all Capra hircus sequenced individuals, European goats and French mainland goats. The best results were obtained with the French goat panel with allele and genotype concordance rates reaching 0.86 and 0.75 in the Alpine and 0.86 and 0.73 in the Saanen breed respectively. Mean correlations tended to be low in both breeds due to the high proportion of variants with low frequencies. For association analysis, imputation was performed using FImpute for 1129 French Alpine and Saanen males using within-breed and French panels on 23,338,436 filtered variants. The association results of both imputation scenarios were then compared. In Saanen goats, a large region on chromosome 19 was significantly linked to semen volume and milk yield in both scenarios. Significant variants for milk yield were annotated for 91 genes on chromosome 19 in Saanen goats. For semen volume, the annotated genes include YBOX2 which is related to azoospermia or oligospermia in other species. New signals for milk yield were detected on chromosome 2 in Alpine goats and on chromosome 5 in Saanen goats when using a multi-breed panel. Conclusion Even with very small reference populations, an acceptable imputation quality can be achieved in French dairy goats. GWAS on imputed sequences confirmed the existence of QTLs and identified new regions of interest in dairy goats. Adding identified candidates to a genotyping array and sequencing more individuals might corroborate the involvement of identified regions while removing potential imputation errors.

Background Random regression models (RRM) are widely used to analyze longitudinal data in genetic evaluation systems because they can better account for time-course changes in environmental effects and additive genetic values of animals by fitting the test-day (TD) specific effects. Our objective was to implement a random regression model for the evaluation of dairy production traits in French goats. Results The data consisted of milk TD records from 30,186 and 32,256 first lactations of Saanen and Alpine goats. Milk yield, fat yield, protein yield, fat content and protein content were considered. Splines were used to model the environmental factors. The genetic and permanent environmental effects were modeled by the same Legendre polynomials. The goodness-of-fit and the genetic parameters derived from functions of the polynomials of orders 0 to 4 were tested. Results were also compared to those from a lactation model with total milk yield calculated over 250 days and to those of a multiple-trait model that considers performance in six periods throughout lactation as different traits. Genetic parameters were consistent between models. Models with fourth-order Legendre polynomials led to the best fit of the data. In order to reduce complexity, computing time, and interpretation, a rank reduction of the variance covariance matrix was performed using eigenvalue decomposition. With a reduction to rank 2, the first two principal components correctly summarized the genetic variability of milk yield level and persistency, with a correlation close to 0 between them. Conclusions A random regression model was implemented in France to evaluate and select goats for yield traits and persistency, which are independent i.e. no genetic correlation between them, in first lactation. Electronic supplementary material The online version of this article (10.1186/s12711-019-0485-3) contains supplementary material, which is available to authorized users.

Lactation curve shape can affect an animal's health, feed requirements, and milk production throughout the year. We implemented a random regression model for the genetic evaluation of lactation curve shapes of dairy traits in French Alpine goats for their first 3 parities. Milk, fat, and protein yields, fat and protein contents, somatic cell score, and fat/protein ratio were considered. The data consisted of test-day records from 49,849 first lactation Alpine goats during their first 3 lactations. The reference model used a Legendre polynomial of order 2 for each parity to describe the genetic and permanent environmental effects, and was compared with a model that combined the second and third parities. A rank reduction of the variance-covariance matrix was also performed using an eigenvalue decomposition for each parity from the 2 models. Genetic parameters were consistent between the models tested. With a reduction to rank 2 and combining the second and third parities, the first 2 principal components correctly summarized the genetic variability of milk yield level and persistency, with a near-nil correlation between the 2, and with a much shorter computation time than the reference model. A favorable correlation of +0.43 between milk yield persistency and fat/protein ratio persistency at the beginning of the lactation was found from buck estimated breeding values.

In 2017, genomic selection was implemented in French dairy goats using the single-step genomic best linear unbiased prediction (ssGBLUP) method, which assumes that all single nucleotide polymorphisms explain the same fraction of genetic variance. However, ssGBLUP is not suitable for protein content, which is controlled by a major gene, i.e. α s 1 casein. This gene explains about 40% of the genetic variation in protein content. In this study, we evaluated the accuracy of genomic prediction using different genomic methods to include the effect of the α s 1 casein gene. Genomic evaluation for protein content was performed with data from the official genetic evaluation on 2955 animals genotyped with the Illumina goat SNP50 BeadChip, 7202 animals genotyped at the α s 1 casein gene and 6,767,490 phenotyped females. Pedigree-based BLUP was compared with regular unweighted ssGBLUP and with three weighted ssGBLUP methods (WssGBLUP, WssGBLUPMax and WssGBLUPSum), which give weights to SNPs according to their effect on protein content. Two other methods were also used: trait-specific marker-derived relationship matrix (TABLUP) using pre-selected SNPs associated with protein content and gene content based on a multiple-trait genomic model that includes α s 1 casein genotypes. We estimated accuracies of predicted genomic estimated breeding values (GEBV) in two populations of goats (Alpine and Saanen). Accuracies of GEBV with ssGBLUP improved by + 5 to + 7 percent points over accuracies from the pedigree-based BLUP model. With the WssGBLUP methods, SNPs that are located close to the α s 1 casein gene had the biggest weights and contributed substantially to the capture of signals from quantitative trait loci. Improvement in accuracy of genomic predictions using the three weighted ssGBLUP methods delivered up to + 6 percent points of accuracy over ssGBLUP. A similar accuracy was obtained for ssGBLUP and TABLUP considering the 20,000 most important SNPs. Incorporating information on the α s 1 casein genotypes based on the gene content method gave similar results as ssGBLUP. The three weighted ssGBLUP methods were efficient for detecting SNPs associated with protein content and for a better prediction of genomic breeding values than ssGBLUP. They also combined fast computing, simplicity and required ssGBLUP to be run only twice.

La sélection génomique, qui a révolutionné la sélection génétique des bovins laitiers notamment, est désormais envisagée dans d’autres filières animales. Chez les caprins laitiers français, le gain de précision attendu des valeurs génomiques était un des questionnements de la filière en raison de la petite taille de la population de référence disponible (825 mâles et 1945 femelles génotypés sur une puce SNP 50K). Le but de cette étude est de tester différentes techniques d’évaluation génomique afin d’obtenir les évaluations génomiques les plus précises possibles. Une étude de la structure génétique de la population de référence caprine constituée d’animaux de races Saanen et Alpine, a révélé de faibles niveaux de déséquilibre de liaison (0,17 entre deux SNP consécutifs), de consanguinité et de parenté au sein de la population, ce qui n’est pas favorable à une bonne précision des évaluations génomiques. Les méthodes d’évaluations génomiques (GBLUP ou Bayésiennes), basées sur des performances pré-corrigées n’ont pas permis une amélioration significative des précisions des évaluations génomiques pour les caractères évalués en routine (caractères de production, de morphologie et comptages de cellules somatiques). Cependant les évaluations génomiques basées sur les performances propres des femelles ont permis d’obtenir des précisions supérieures à celles obtenues sur ascendance. La sélection génomique est donc envisageable chez les caprins laitiers français. Ces précisions peuvent également être légèrement augmentées par l’inclusion de gènes majeurs tels que celui de la caséine αs1., Genomic selection, which is revolutionizing genetic selection in dairy cattle is now considered in the breeding of other animals. In French dairy goats, gain in accuracy using genomic selection has been questionned due to the small size of the reference population (825 males and 1 945 females genotyped). The aim of this study was to investigate how to reach adequate genomic evaluation accuracy in the French dairy goat population. The study of a reference population structure (Alpine and Saanen breeds) showed that the level of linkage disequilibrium (0.17 between two consecutive SNP), inbreeding and the relationship between the reference and candidate population were not ideal to maximize genomic evaluation accuracy. Two steps genomic evaluation (GBLUP, Bayesian) based on performances corrected for fixed effects did not improve genetic evaluation accuracy compared to classical evaluations for milk production traits, udder type traits and somatic cells scores. When using genomic evaluation directly based on female performances (single step), accuracy of genomic evaluation is higher than the accuracy level obtained from the ascendance. Genomic selection is feasible in French dairy goats. Accuracies could be slightly improved integrating a major gene such as αs1 casein.

Ovarian folliculogenesis corresponds to the development of follicles leading to either ovulation or degeneration, this latter process being called atresia. Even if atresia involves apoptosis, its mechanism is not well-understood. The objective of this project was to analyse global gene expression in pig granulosa cells of ovarian follicles during atresia. The transcriptome analysis was performed using 9216 cDNAs microarray to identify gene networks and candidate genes involved in pig ovarian follicular atresia. One thousand six hundred and eighty four significantly regulated genes were differentially regulated between small healthy follicles and small atretic follicles. Among them, two hundred and eighty seven genes had a fold-change higher than 2 between the two follicle groups. Eleven genes (DKK3, GADD45A, CAMTA2, CCDC80, DAPK2, ECSIT, MSMB, NUPR1, RUNX2, SAMD4A, and ZNF628) having a fold-change higher than 5 between groups could likely serve as markers of follicular atresia. Moreover, automatic confrontation of deregulated genes with literature data enlightened 93 genes as regulatory candidates of pig granulosa cell atresia. Among these genes known to be inhibitors of apoptosis, stimulators of apoptosis or tumor suppressors INHBB, HNF4, CLU, different interleukins (IL5, IL24), TNF-associated receptor (TNFR1), and cytochrome-c oxidase (COX) were suggested as playing an important role in porcine atresia. Present study also enlists key upstream regulators in follicle atresia based on our results and on a literature review. The novel gene candidates and gene networks identified in the current study lead to a better understanding of the molecular regulation of ovarian follicular atresia.

Genomic best linear unbiased prediction methods assume that all markers explain the same fraction of the genetic variance and do not account effectively for genes with major effects such as the α s1 casein polymorphism in dairy goats. In this study, we investigated methods to include the available α s1 casein genotype effect in genomic evaluations of French dairy goats. First, the α s1 casein genotype was included as a fixed effect in genomic evaluation models based only on bucks that were genotyped at the α s1 casein locus. Less than 1 % of the females with phenotypes were genotyped at the α s1 casein gene. Thus, to incorporate these female phenotypes in the genomic evaluation, two methods that allowed for this large number of missing α s1 casein genotypes were investigated. Probabilities for each possible α s1 casein genotype were first estimated for each female of unknown genotype based on iterative peeling equations. The second method is based on a multiallelic gene content approach. For each model tested, we used three datasets each divided into a training and a validation set: (1) two-breed population (Alpine + Saanen), (2) Alpine population, and (3) Saanen population. The α s1 casein genotype had a significant effect on milk yield, fat content and protein content. Including an α s1 casein effect in genetic and genomic evaluations based only on male known α s1 casein genotypes improved accuracies (from 6 to 27 %). In genomic evaluations based on all female phenotypes, the gene content approach performed better than the other tested methods but the improvement in accuracy was only slightly better (from 1 to 14 %) than that of a genomic model without the α s1 casein effect. Including the α s1 casein effect in a genomic evaluation model for French dairy goats is possible and useful to improve accuracy. Difficulties in predicting the genotypes for ungenotyped animals limited the improvement in accuracy of the obtained estimated breeding values.

Après une présentation du contexte de la sélection ovine laitière française et de la stratégie de sélection génomique qui en découle, cet article retrace l’ensemble des travaux de recherche et développement entrepris depuis 2009 pour mettre en œuvre cette stratégie et qui débouchent sur le déploiement de la sélection génomique en 2015. Sont abordés successivement la constitution des populations de référence, les investissements en méthodes et stratégies de calcul d’index génomique, les résultats de précision des index génomiques en ovins laitiers, l’expérimentation génomique originale mise en œuvre en race Lacaune, la conception, la modélisation et l’optimisation de schémas de sélection génomique adaptés aux spécificités des races ovines laitières, les choix opérés par les entreprises et organismes de sélection et leurs conséquences organisationnelles et contractuelles. Les perspectives ouvertes par la sélection génomique sont évoquées : inclusion de nouveaux caractères, plus grande flexibilité, meilleure gestion de la variabilité génétique. La sélection génomique est effective depuis 2015 en race Lacaune et devrait être déployée en 2017 dans les races pyrénéennes., After a presentation of the background of French dairy sheep selection and the resulting strategy of genomic selection, this article considers the different research x development achievements undertaken since 2009 in order to assess the feasibility of this strategy and which results in the implementation of genomic selection in 2015. The following topics are successively reviewed: the constitution of the reference populations, the methods and strategies of genomic evaluation, the accuracy obtained for genomic prediction in the different dairy sheep breeds, the original experiment conducted in the Lacaune breed, the conception, modeling and optimization of dairy-sheep-customized genomic breeding schemes, the decisions made by the breeding societies and their organizational and contractual implications. The main expectations opened by this new field of genomic selection are highlighted: selection for new traits, increased flexibility in the breeding programs, more efficient management of genetic variability. The Lacaune breed has been switched to genomic selection in 2015, whereas in the Pyrenean breeds, breeders intend to do it by 2017.

Some dairy farmers opt to omit one milking, either incidentally or weekly, without changing other milking times. This practice entails an extended milking interval of 24 h (24h-MI), which is associated with a decrease in milk yield. This decrease varies among cows and could be partly due to factors such as stage of lactation and milk yield level. The aim of this study was to describe the average and individual responses in terms of loss and carryover effects of a 24h-MI on milk yield. The influence of factors such as parity, stage of lactation, and milk yield potential were investigated, together with response repeatability. Our trial used 292 Holstein-Friesian cows, and consisted of 3 successive periods: 1 wk of twice-daily milking (TDM) as a control, one 24h-MI, and then 13 d of TDM. The number of observations per cow ranged from 1 to 9, with no more than three 24h-MI per lactation. The 24h-MI reduced milk yield by 23% (7.8 kg on average) and milk lactose content by 2.6 g/kg on the 24h-MI day. Milk fat and protein content, and somatic cell score increased by 3.0 g/kg, 0.5 g/kg, and 0.4 units, respectively. No significant carryover effect was found of a 24h-MI on milk yield or milk composition 2 wk after resumption of TDM. Milk yield loss and recovery varied widely (coefficient of variation 62%), and the relationship between milk loss and milk recovery showed substantial variation (residual standard deviation 2.1 kg/d). Cows with a greater milk potential level lost more milk yield but recovered more milk, with no influence on recovery:loss ratio. Cows in early lactation recovered the lost milk yield faster. Repeatability of the responses to a 24h-MI was 44% for milk yield loss (kg/d), 57% for relative milk yield loss (%), 33% for milk yield recovery (kg/d), and 0% for milk recovery:loss ratio (%), suggesting a genetically determined ability to limit loss when one milking is omitted. To conclude, a 24h-MI caused higher milk yield losses than reported in previous studies. Stage of lactation, estimated potential milk yield level, and parity explained the cows' response to the 24h-MI, but did not account for all the individual variability.

Avec l’arrivée de données de génotypage à haut débit, il est maintenant possible d’estimer la valeur génétique d’animaux candidats à la sélection dès leur naissance, sans attendre la collecte de phénotypes. La sélection génomique bouleverse complètement les perspectivesen amélioration génétique. Elle nécessite la constitution d’une population de référence formée d’animaux génotypés (jusqu’à récemment, il s’agissait principalement de mâles) et ayant des performances précises, par exemple la performance moyenne de leurs filles. Les évaluations génomiques consistent à prédire les phénotypes dans cette population de référence comme la somme des effets des marqueurs moléculaires. Le problème méthodologique principal est que le nombre d’effets à estimer est typiquement beaucoup plus élevé que le nombre de phénotypes disponibles. Nous décrivons les idées générales de diverses familles de méthodes proposées : BLUP génomique basé sur une parenté entre individus calculée à partir des marqueurs, méthodes Bayésiennes plus flexibles mais aussi plus coûteuses, méthodes de sélection de variables, méthode en une seule étape qui combine évaluation génétique nationale et évaluation génomique. La précision des évaluations génomiques est faite par validation croisée chez les animaux les plus jeunes de la population de référence. La taille de la population de référence, la manière de prendre en compte les QTL à effet fort et le degré d’apparentement entre candidats à la sélection et animaux de la population de référence ont un impact non négligeable sur l’efficacité desméthodes de sélection génomique., With the advent of high throughput genotyping, it is now possible to estimate breeding values of selection candidates at birthwithout waiting for phenotypic data collection. Genomic selection is deeply changing future prospects in genetic improvement. Itrequires the creation of a reference population made of genotyped animals (generally males) with precise phenotypes, such as averagedaughter performances. Genomic evaluations consist in predicting phenotypes in this reference population as the sum of molecularmarker effects. The main methodological challenge is the large number of effects to estimate, usually much larger than thenumber of available phenotypes. We briefly describe the various families of proposed methods: genomic BLUP based on relationships computed from marker information, Bayesian methods which are more flexible but more tedious, variable selection methods and asingle step method which combines national genetic evaluation and genomic evaluation. The precision of genomic selection is donevia cross validation among the youngest animals of the reference population. The size of the reference population, the way QTL withlarge effects are modeled, the relationship between selection candidates and the reference population have a significant impact on theefficiency of genomic selection methods.

To understand the mechanisms underlying milk ability and more precisely the kinetics of milk emission, we compared teat transcriptome profiles from Lacaune ewes in the tails of the milk flow phenotypic distribution. Two different arrays containing respectively 1896 and 13 168 PCR products selected from several tissue-specific cDNA libraries, including mammary gland, allowed the identification of 73 differentially expressed genes between teats from high and low milk flow ewes. Genes involved in muscle contraction were identified as over-expressed, and genes encoding collagen were found to be under-expressed in teats from low milk flow ewes. We confirmed this underexpression of COL1A1 and COL1A2 in low-milk flow ewes using RT-qPCR. These results suggest that milking ability may be due to the capacity of the teat sphincter to relax during mechanical milking. We propose that an optimal condition for mechanical milking may require proper relaxation of the teats. To our knowledge, this is the first transcriptomic analysis studying milking ability, using udder tissue for gene expression profiling, which demonstrates that mechanical milking ability is not only determined by morphological features but also by tissue composition.

Although susceptibility to scrapie is largely controlled by the PrP gene, the role of other genes that affect scrapie resistance in sheep is now confirmed. Following the detection of quantitative trait loci (QTL) on chromosomes 6 and 18 in a half-sib family with an ARQ/VRQ susceptible PrP genotype, the whole pedigree of a naturally infected flock was investigated to confirm these QTL regions in different PrP genotypes. The present study has allowed us to confirm the QTL on chromosome 18, and to demonstrate the QTL effects in several PrP genotypes.

The aim of this paper was to describe, and when possible compare, the multivariate methods used by the participants in the EADGENE WP1.4 workshop. The first approach was for class discovery and class prediction using evidence from the data at hand. Several teams used hierarchical clustering (HC) or principal component analysis (PCA) to identify groups of differentially expressed genes with a similar expression pattern over time points and infective agent (E. coli or S. aureus). The main result from these analyses was that HC and PCA were able to separate tissue samples taken at 24 h following E. coli infection from the other samples. The second approach identified groups of differentially co-expressed genes, by identifying clusters of genes highly correlated when animals were infected with E. coli but not correlated more than expected by chance when the infective pathogen was S. aureus. The third approach looked at differential expression of predefined gene sets. Gene sets were defined based on information retrieved from biological databases such as Gene Ontology. Based on these annotation sources the teams used either the GlobalTest or the Fisher exact test to identify differentially expressed gene sets. The main result from these analyses was that gene sets involved in immune defence responses were differentially expressed.

Microarrays allow researchers to measure the expression of thousands of genes in a single experiment. Before statistical comparisons can be made, the data must be assessed for quality and normalisation procedures must be applied, of which many have been proposed. Methods of comparing the normalised data are also abundant, and no clear consensus has yet been reached. The purpose of this paper was to compare those methods used by the EADGENE network on a very noisy simulated data set. With the a priori knowledge of which genes are differentially expressed, it is possible to compare the success of each approach quantitatively. Use of an intensity-dependent normalisation procedure was common, as was correction for multiple testing. Most variety in performance resulted from differing approaches to data quality and the use of different statistical tests. Very few of the methods used any kind of background correction. A number of approaches achieved a success rate of 95% or above, with relatively small numbers of false positives and negatives. Applying stringent spot selection criteria and elimination of data did not improve the false positive rate and greatly increased the false negative rate. However, most approaches performed well, and it is encouraging that widely available techniques can achieve such good results on a very noisy data set.

Animal nutrition considerably affects milk composition that influences its nutritional quality. Milk component synthesis and secretion by the mammary gland involve expression of a large number of genes whose nutritional regulation remains poorly defined. In this study, we examined the effect of food deprivation (FD) on the expression of 8379 genes in caprine mammary gland using a bovine oligonucleotide microarray. Twelve lactating goats were assigned to 2 groups based on their feeding level (control diet ad libitum vs. 48-h FD). We identified 161 genes whose expression was altered by FD. Most of these genes (88%) were downregulated, suggesting a stress response by the mammary gland. In particular, the decrease in expression of genes involved in milk protein, lactose, and lipid metabolism could contribute together with the shortage of nutrients to the drop in milk protein, lactose, and fat secretion. In addition, this study highlights modification of the expression of at least 14 genes that could be responsible for a slowdown in mammary cell proliferation and differentiation and/or an increase in programmed cell death in response to 48-h FD in goats.

This study presents a dynamic characterization of the sheep milk transcriptome aiming at achieving a better understanding of the sheep lactating mammary gland. Transcriptome sequencing (RNA-seq) was performed on total RNA extracted from milk somatic cells from ewes on days 10, 50, 120 and 150 after lambing. The experiment was performed in Spanish Churra and Assaf breeds, which differ in their milk production traits. Nearly 67% of the annotated genes in the reference genome (Oar_v3.1) were expressed in ovine milk somatic cells. For the two breeds and across the four lactation stages studied, the most highly expressed genes encoded caseins and whey proteins. We detected 573 differentially expressed genes (DEGs) across lactation points, with the largest differences being found, between day 10 and day 150. Upregulated GO terms at late lactation stages were linked mainly to developmental processes linked to extracellular matrix remodeling. A total of 256 annotated DEGs were detected in the Assaf and Churra comparison. Some genes selectively upregulated in the Churra breed grouped under the endopeptidase and channel activity GO terms. These genes could be related to the higher cheese yield of this breed. Overall, this study provides the first integrated overview on sheep milk gene expression.

International audience; Background: Haemonchosis is a parasitic disease that causes severe economic losses in sheep industry. In recent years, the increasing resistance of the parasite to anthelmintics has raised the need for alternative control strategies. Genetic selection is a promising alternative but its efficacy depends on the availability of genetic variation and on the occurrence of favourable genetic correlations between the traits included in the breeding goal. The objective of this study was twofold. First, to estimate both the heritability of and the genetic correlations between growth traits and parasite resistance traits, using bivariate linear mixed animal models, from the phenotypes and genotypes of 1004 backcross lambs (considered as a single population), which underwent two subsequent experimental infestations protocols with Haemonchus contortus. Second, to compare the precision of the estimates when using two different relationship matrices: including pedigree information only or including also SNP (single nucleotide polymorphism) information.Results: Heritabilities were low for average daily gain before infestation (0.10 to 0.15) and average daily gain during the first infestation (0.11 to 0.16), moderate for faecal egg counts during the first infestation (0.21 to 0.38) and faecal egg counts during the second infestation (0.48 to 0.55). Genetic correlations between both growth traits and faecal egg count during the naive infestation were equal to zero but the genetic correlation between faecal egg count during the second infestation and growth was positive in a Haemonchus contortus free environment and negative in a contaminated environment. The standard errors of the estimates obtained by including SNP information were smaller than those obtained by including pedigree information only.Conclusions: The genetic parameters estimates suggest that growth performance can be selected for independently of selection on resistance to naive infestation. Selection for increased growth in a non-contaminated environment could lead to more susceptible animals with long-term exposure to the infestation but it could be possible to select for increased growth in a contaminated environment while also increasing resistance to the long-term exposure to the parasite. The use of molecular information increases the precision of the estimates.

International audience; Background: All progeny-tested bucks from the two main French dairy goat breeds (Alpine and Saanen) were genotyped with the Illumina goat SNP50 BeadChip. The reference population consisted of 677 bucks and 148 selection candidates. With the two-step approach based on genomic best linear unbiased prediction (GBLUP), prediction accuracy of candidates did not outperform that of the parental average. We investigated a GBLUP method based on a single-step approach, with or without blending of the two breeds in the reference population.Methods: Three models were used: (1) a multi-breed model, in which Alpine and Saanen breeds were considered as a single breed; (2) a within-breed model, with separate genomic evaluation per breed; and (3) a multiple-trait model, in which a trait in the Alpine was assumed to be correlated to the same trait in the Saanen breed, using three levels of between-breed genetic correlations (ρ): ρ = 0, ρ = 0.99, or estimated ρ. Quality of genomic predictions was assessed on progeny-tested bucks, by cross-validation of the Pearson correlation coefficients for validation accuracy and the regression coefficients of daughter yield deviations (DYD) on genomic breeding values (GEBV). Model-based estimates of average accuracy were calculated on the 148 candidates.Results: The genetic correlations between Alpine and Saanen breeds were highest for udder type traits, ranging from 0.45 to 0.76. Pearson correlations with the single-step approach were higher than previously reported with a two-step approach. Correlations between GEBV and DYD were similar for the three models (within-breed, multi-breed and multiple traits). Regression coefficients of DYD on GEBV were greater with the within-breed model and multiple-trait model with ρ = 0.99 than with the other models. The single-step approach improved prediction accuracy of candidates from 22 to 37% for both breeds compared to the two-step method.Conclusions: Using a single-step approach with GBLUP, prediction accuracy of candidates was greater than that based on parent average of official evaluations and accuracies obtained with a two-step approach. Except for regression coefficients of DYD on GEBV, there were no significant differences between the three models.

Genomic selection in Lacaune dairy sheep was investigated based on genotypes from the OvineSNP50 BeadChip (Illumina Inc., San Diego, CA). Historical artificial insemination progeny-tested rams formed a population of 2,892 genotyped rams. Additional ungenotyped rams and females were included by single-step genomic BLUP (ssGBLUP). Three prediction strategies were tried: pseudo-BLUP (using all rams and daughter yield deviations), pseudo-ssGBLUP (using all rams and daughter yield deviations), and regular ssGBLUP (using all phenotypes and pedigree in an animal model). The population linkage disequilibrium was determined, with an average squared correlation coefficient of 0.11 for markers closer than 0.1cM (lower than in dairy cattle). The estimated effective population is 370 individuals. Gain in accuracy of genomic selection over parent averages ranged from 0.10 to 0.20. Highest accuracies and lowest bias were found using regular ssGBLUP. Transition to a genomic breeding scheme is possible but costs need to be carefully evaluated.

This article presents an extension of the methodology developed by Gilmour et al. [19], for ordered categorical data, taking into account the heterogeneity of residual variances of latent variables. Heterogeneity of residual variances is described via a structural linear model on log-variances. This method involves two main steps: i) a ’marginalization’ with respect to the random effects leading to quasi-score estimators; ii) an approximation of the variance-covariance matrix of the observations which leads to an analogue of the Henderson mixed model equations for continuous Gaussian data. This methodology is illustrated by a numerical example of footshape in sheep., Cet article présente une extension de la méthodologie développée par Gilmour et al. [19] dans le cas de variables qualitatives ordonnées, prenant en compte l’hétérogénéité des variances résiduelles des variables latentes. L’hétérogénéité des variances résiduelles est décrite par un modèle linéaire structurel sur les logarithmes des variances. Cette méthode comprend deux étapes principales : i) une « marginalisation » par rapport aux effets aléatoires qui conduit, grâce aux équations de quasi-score, à l’estimation des paramètres ; ii) une approximation de la matrice de variance-covariance des observations qui aboutit à un système analogue aux équations du modèle mixte d’Henderson dans le cas de variables continues gaussiennnes. Cette méthodologie est illustrée par un exemple sur la forme des pieds chez le mouton.

Recently, the amount of available single nucleotide polymorphism (SNP) marker data has considerably increased in dairy cattle breeds, both for research purposes and for application in commercial breeding and selection programs. Bayesian methods are currently used in the genomic evaluation of dairy cattle to handle very large sets of explanatory variables with a limited number of observations. In this study, we applied 2 Bayesian methods, BayesCπ and Bayesian least absolute shrinkage and selection operator (LASSO), to 2 genotyped and phenotyped reference populations consisting of 3,940 Holstein bulls and 1,172 Montbeliarde bulls with approximately 40,000 polymorphic SNP. We compared the accuracy of the Bayesian methods for the prediction of 3 traits (milk yield, fat content, and conception rate) with pedigree-based BLUP, genomic BLUP, partial least squares (PLS) regression, and sparse PLS regression, a variable selection PLS variant. The results showed that the correlations between observed and predicted phenotypes were similar in BayesCπ (including or not pedigree information) and Bayesian LASSO for most of the traits and whatever the breed. In the Holstein breed, Bayesian methods led to higher correlations than other approaches for fat content and were similar to genomic BLUP for milk yield and to genomic BLUP and PLS regression for the conception rate. In the Montbeliarde breed, no method dominated the others, except BayesCπ for fat content. The better performances of the Bayesian methods for fat content in Holstein and Montbeliarde breeds are probably due to the effect of the DGAT1 gene. The SNP identified by the BayesCπ, Bayesian LASSO, and sparse PLS regression methods, based on their effect on the different traits of interest, were located at almost the same position on the genome. As the Bayesian methods resulted in regressions of direct genomic values on daughter trait deviations closer to 1 than for the other methods tested in this study, Bayesian methods are suggested for genomic evaluations of French dairy cattle.

National audience; L’omission d’une traite par jour entraîne, chez la vache laitière, une importante perte de production suivie d’un effet rémanent négatif lors du retour à 2 traites par jour, après 3 semaines de monotraite. La perte de production lors du passage à 1 traite/j est très variable entre animaux (Rémond et Pomiès, 2005), rendant difficile une prévision. L’identification, et plus encore la sélection d’animaux aptes à supporter la monotraite nécessite de mesurer un phénotype corrélé aux 2 aspects de cette aptitude (perte, effet rémanent), facilement mesurable en ferme et assez tôt dans la lactation. La présente étude a pour objet de décrire la variabilité entre individus des 2 aspects de l’aptitude à la monotraite, et aussi d’identifier des premiers phénotypes prédicteurs

Genomic selection involves computing a prediction equation from the estimated effects of a large number of DNA markers based on a limited number of genotyped animals with phenotypes. The number of observations is much smaller than the number of independent variables, and the challenge is to find methods that perform well in this context. Partial least squares regression (PLS) and sparse PLS were used with a reference population of 3,940 genotyped and phenotyped French Holstein bulls and 39,738 polymorphic single nucleotide polymorphism markers. Partial least squares regression reduces the number of variables by projecting independent variables onto latent structures. Sparse PLS combines variable selection and modeling in a one-step procedure. Correlations between observed phenotypes and phenotypes predicted by PLS and sparse PLS were similar, but sparse PLS highlighted some genome regions more clearly. Both PLS and sparse PLS were more accurate than pedigree-based BLUP and generally provided lower correlations between observed and predicted phenotypes than did genomic BLUP. Furthermore, PLS and sparse PLS required similar computing time to genomic BLUP for the study of 6 traits.

Chantier qualité GA; International audience; Goat milk somatic cell counts have been collected for several years in France by the national milk recording organization. Information is used for health management, because repeatedly elevated somatic cell counts are a good indirect predictor of intramammary infection. Genetic parameters were estimated for 67,882 and 49,709 primiparous goats of the dairy Alpine and Saanen breeds, respectively, with complete information for milk somatic cell counts and milk production traits. About 40% of the goats had additional information for 11 udder type traits scored by official classifiers of the breeders’ association CAPGENES. Estimates were obtained by REML with an animal model. The studied trait was lactation somatic cell score (LSCS), the weighted mean of somatic cell score (log-transformed SCC) adjusted for lactation stage. Heritability of LSCS was 0.20 and 0.24 in the Alpine and Saanen breeds, respectively. Relationships with milk production and udder type traits were additionally estimated by using multitrait analyses. Heritability estimates in first lactation ranged from 0.30 to 0.35 for lactation milk, fat, and protein yields; from 0.60 to 0.67 for fat and protein contents; and from 0.22 to 0.50 for udder type traits. Genetic correlations of somatic cell score with milk production traits were generally low, ranging from −0.13 to 0.12. Slightly more negative correlations were estimated for fat content: −0.18 and −0.20 in Saanen and Alpine breeds, respectively. Lactation somatic cell score was genetically correlated with udder floor position (rg=−0.24 and −0.19 in the Alpine and Saanen breeds, respectively), and, in Saanen, teat length, teat width, and teat form (rg=0.29, 0.34 and −0.27, respectively). These results suggest that a reduction in somatic cell count can be achieved by selection while still improving milk production and udder type and teat traits.

SummaryEmpirical experience with genomic selection in dairy cattle suggests that the distribution of the effects of single nucleotide polymorphisms (SNPs) might be far from normality for some traits. An alternative, avoiding the use of arbitrary prior information, is the Bayesian Lasso (BL). Regular BL uses a common variance parameter for residual and SNP effects (BL1Var). We propose here a BL with different residual and SNP effect variances (BL2Var), equivalent to the original Lasso formulation. The λ parameter in Lasso is related to genetic variation in the population. We also suggest precomputing individual variances of SNP effects by BL2Var, to be later used in a linear mixed model (HetVar-GBLUP). Models were tested in a cross-validation design including 1756 Holstein and 678 Montbéliarde French bulls, with 1216 and 451 bulls used as training data; 51 325 and 49 625 polymorphic SNP were used. Milk production traits were tested. Other methods tested included linear mixed models using variances inferred from pedigree estimates or integrated out from the data. Estimates of genetic variation in the population were close to pedigree estimates in BL2Var but not in BL1Var. BL1Var shrank breeding values too little because of the common variance. BL2Var was the most accurate method for prediction and accommodated well major genes, in particular for fat percentage. BL1Var was the least accurate. HetVar-GBLUP was almost as accurate as BL2Var and allows for simple computations and extensions.

Axis specification in mouse is determined by a sequence of reciprocal interactions between embryonic and extra-embryonic tissues so that a few extra-embryonic genes appear as ‘patterning’ the embryo. Considering these interactions as essential, but lacking in most mammals the genetically driven approaches used in mouse and the corresponding patterning mutants, we examined whether a molecular signature originating from extra-embryonic tissues could relate to the developmental stage of the embryo proper and predict it. To this end, we have profiled bovine extra-embryonic tissues at peri-implantation stages, when gastrulation and early neurulation occur, and analysed the subsequent expression profiles through the use of predictive methods as previously reported for tumour classification. A set of six genes (CALM1,CPA3,CITED1,DLD,HNRNPDL, andTGFB3), half of which had not been previously associated with any extra-embryonic feature, appeared significantly discriminative and mainly dependent on embryonic tissues for its faithful expression. The predictive value of this set of genes for gastrulation and early neurulation stages, as assessed on naive samples, was remarkably high (93%).In silicoconnected to the bovine orthologues of the mouse patterning genes, this gene set is proposed as a new trait for embryo staging. As such, this will allow saving the bovine embryo proper for molecular or cellular studies. To us, it offers as well new perspectives for developmental phenotyping and modelling of embryonic/extra-embryonic co-differentiation.

SummaryFor genomic selection methods, the statistical challenge is to estimate the effect of each of the available single-nucleotide polymorphism (SNP). In a context where the number of SNPs (p) is much higher than the number of bulls (n), this task may lead to a poor estimation of these SNP effects if, as for genomic BLUP (gBLUP), all SNPs have a non-null effect. An alternative is to use approaches that have been developed specifically to solve the ‘p>>n’ problem. This is the case of variable selection methods and among them, we focus on the Elastic-Net (EN) algorithm that is a penalized regression approach. Performances of EN, gBLUP and pedigree-based BLUP were compared with data from three French dairy cattle breeds, giving very encouraging results for EN. We tried to push further the idea of improving SNP effect estimates by considering fewer of them. This variable selection strategy was considered both in the case of gBLUP and EN by adding an SNP pre-selection step based on quantitative trait locus (QTL) detection. Similar results were observed with or without a pre-selection step, in terms of correlations between direct genomic value (DGV) and observed daughter yield deviation in a validation data set. However, when applied to the EN algorithm, this strategy led to a substantial reduction of the number of SNPs included in the prediction equation. In a context where the number of genotyped animals and the number of SNPs gets larger and larger, SNP pre-selection strongly alleviates computing requirements and ensures that national evaluations can be completed within a reasonable time frame.

Background The existence of a genetic basis for host responses to bacterial intramammary infections has been widely documented, but the underlying mechanisms and the genes are still largely unknown. Previously, two divergent lines of sheep selected for high/low milk somatic cell scores have been shown to be respectively susceptible and resistant to intramammary infections by Staphylococcus spp. Transcriptional profiling with an 15K ovine-specific microarray of the milk somatic cells of susceptible and resistant sheep infected successively by S. epidermidis and S. aureus was performed in order to enhance our understanding of the molecular and cellular events associated with mastitis resistance. Results The bacteriological titre was lower in the resistant than in the susceptible animals in the 48 hours following inoculation, although milk somatic cell concentration was similar. Gene expression was analysed in milk somatic cells, mainly represented by neutrophils, collected 12 hours post-challenge. A high number of differentially expressed genes between the two challenges indicated that more T cells are recruited upon inoculation by S. aureus than S. epidermidis. A total of 52 genes were significantly differentially expressed between the resistant and susceptible animals. Further Gene Ontology analysis indicated that differentially expressed genes were associated with immune and inflammatory responses, leukocyte adhesion, cell migration, and signal transduction. Close biological relationships could be established between most genes using gene network analysis. Furthermore, gene expression suggests that the cell turn-over, as a consequence of apoptosis/granulopoiesis, may be enhanced in the resistant line when compared to the susceptible line. Conclusions Gene profiling in resistant and susceptible lines has provided good candidates for mapping the biological pathways and genes underlying genetically determined resistance and susceptibility towards Staphylococcus infections, and opens new fields for further investigation.

National audience

International audience; A divergent selection experiment in sheep was implemented to study the consequences of log-transformed somatic cell score (SCS)-based selection on resistance to natural intramammary infections. Using dams and progeny-tested rams selected for extreme breeding values for SCS, we created 2 groups of ewes with a strong divergence in SCS of approximately 3 genetic standard deviations. A survey of 84 first-lactation ewes of both the High and Low SCS lines indicated favorable responses to SCS-based selection on resistance to both clinical and subclinical mastitis. All clinical cases (n=5) occurred in the High SCS line. Additionally, the frequency of chronic clinical mastitis, as detected by the presence of parenchymal abscesses, was much greater in the High SCS line (n=21) than in the Low SCS line (n=1). According to monthly milk bacteriological examinations of udder halves, the prevalence of infection was significantly greater (odds ratio = 3.1) in the High SCS line than in the Low SCS line, with predicted probabilities of 37 and 16%, respectively. The most frequently isolated bacteria responsible for mastitis were staphylococci: Staphylococcus auricularis (42.6% of positive samples), Staphylococcus simulans, Staphylococcus haemoliticus, Staphylococcus xylosus, Staphylococcus chromogenes, Staphylococcus lentus, Staphylococcus warneri, and Staphylococcus aureus. The incidence of positive bacteriology was greater in the High SCS line (39%) than in the Low SCS line (12%) at lambing, indicating that High SCS line ewes were especially susceptible to postpartum subclinical mastitis. Negativation of bacteriological results from one sampling time point to the next was markedly different between lines after weaning (e.g., 41 and 84% in the High and Low SCS lines, respectively). This result was consistent with differences in the duration of infection, which was much greater in the High SCS line compared with the Low SCS line. Finally, ewes from the High SCS line consistently had greater SCS in positive milk samples than did ewes from the Low SCS line (+2.04 SCS, on average), with an especially large difference between lines during the suckling period (+3.42 SCS). Altogether, the preliminary results suggest that the better resistance of Low SCS line ewes, compared with High SCS line ewes, was principally characterized by a better ability to limit infections during the peripartum period, to eliminate infections during lactation, and quantitatively to limit the inflammation process and its clinical consequences.

Background As presented in the introduction paper, three sets of differentially regulated genes were found after the analysis of the chicken infection data set from EADGENE. Different methods were used to interpret these results. Results GOTM, Pathway Studio and Ingenuity softwares were used to investigate the three lists of genes. The three softwares allowed the analysis of the data and highlighted different networks. However, only one set of genes, showing a differential expression between primary and secondary response gave significant biological interpretation. Conclusion Combining these databases that were developed independently on different annotation sources supplies a useful tool for a global biological interpretation of microarray data, even if they may contain some imperfections (e.g. gene not or not well annotated).

Background The aim of this work was to study the performances of 2 predictive statistical tools on a data set that was given to all participants of the Eadgene-SABRE Post Analyses Working Group, namely the Pig data set of Hazard et al. (2008). The data consisted of 3686 gene expressions measured on 24 animals partitioned in 2 genotypes and 2 treatments. The objective was to find biomarkers that characterized the genotypes and the treatments in the whole set of genes. Methods We first considered the Random Forest approach that enables the selection of predictive variables. We then compared the classical Partial Least Squares regression (PLS) with a novel approach called sparse PLS, a variant of PLS that adapts lasso penalization and allows for the selection of a subset of variables. Results All methods performed well on this data set. The sparse PLS outperformed the PLS in terms of prediction performance and improved the interpretability of the results. Conclusion We recommend the use of machine learning methods such as Random Forest and multivariate methods such as sparse PLS for prediction purposes. Both approaches are well adapted to transcriptomic data where the number of features is much greater than the number of individuals.

Background In the context of systems biology, few sparse approaches have been proposed so far to integrate several data sets. It is however an important and fundamental issue that will be widely encountered in post genomic studies, when simultaneously analyzing transcriptomics, proteomics and metabolomics data using different platforms, so as to understand the mutual interactions between the different data sets. In this high dimensional setting, variable selection is crucial to give interpretable results. We focus on a sparse Partial Least Squares approach (sPLS) to handle two-block data sets, where the relationship between the two types of variables is known to be symmetric. Sparse PLS has been developed either for a regression or a canonical correlation framework and includes a built-in procedure to select variables while integrating data. To illustrate the canonical mode approach, we analyzed the NCI60 data sets, where two different platforms (cDNA and Affymetrix chips) were used to study the transcriptome of sixty cancer cell lines. Results We compare the results obtained with two other sparse or related canonical correlation approaches: CCA with Elastic Net penalization (CCA-EN) and Co-Inertia Analysis (CIA). The latter does not include a built-in procedure for variable selection and requires a two-step analysis. We stress the lack of statistical criteria to evaluate canonical correlation methods, which makes biological interpretation absolutely necessary to compare the different gene selections. We also propose comprehensive graphical representations of both samples and variables to facilitate the interpretation of the results. Conclusion sPLS and CCA-EN selected highly relevant genes and complementary findings from the two data sets, which enabled a detailed understanding of the molecular characteristics of several groups of cell lines. These two approaches were found to bring similar results, although they highlighted the same phenomenons with a different priority. They outperformed CIA that tended to select redundant information.

From EADGENE and SABRE Post-analyses Workshop Lelystad, The Netherlands. 12–14 November 2008 Background: The aim of this paper was to describe and compare the methods used and the results obtained by the participants in a joint EADGENE (European Animal Disease Genomic Network of Excellence) and SABRE (Cutting Edge Genomics for Sustainable Animal Breeding) workshop focusing on post analysis of microarray data. The participating groups were provided with identical lists of microarray probes, including test statistics for three different contrasts, and the normalised log-ratios for each array, to be used as the starting point for interpreting the affected probes. The data originated from a microarray experiment conducted to study the host reactions in broilers occurring shortly after a secondary challenge with either a homologous or heterologous species of Eimeria. Results: Several conceptually different analytical approaches, using both commercial and public available software, were applied by the participating groups. The following tools were used: Ingenuity Pathway Analysis, MAPPFinder, LIMMA, GOstats, GOEAST, GOTM, Globaltest, TopGO, ArrayUnlock, Pathway Studio, GIST and AnnotationDbi. The main focus of the approaches was to utilise the relation between probes/genes and their gene ontology and pathways to interpret the affected probes/genes. The lack of a well-annotated chicken genome did though limit the possibilities to fully explore the tools. The main results from these analyses showed that the biological interpretation is highly dependent on the statistical method used but that some common biological conclusions could be reached. Conclusion: It is highly recommended to test different analytical methods on the same data set and compare the results to obtain a reliable biological interpretation of the affected genes in a DNA microarray experiment.

Ovarian antral follicular development is clearly dependent on pituitary gonadotrophins FSH and LH. Although the endocrine mechanism that controls ovarian folliculogenesis leading to ovulation is quite well understood, the detailed mechanisms and molecular determinants in the different follicular compartments remain to be clarified. The aim of this study was to identify the genes differentially expressed in pig granulosa cells along the terminal ovarian follicle growth, to gain a comprehensive view of these molecular mechanisms. First, we developed a specific micro-array using cDNAs from suppression subtractive hybridization libraries (345 contigs) obtained by comparison of three follicle size classes: small, medium and large antral healthy follicles. In a second step, a transcriptomic analysis using cDNA probes from these three follicle classes identified 79 differentially expressed transcripts along the terminal follicular growth and 26 predictive genes of size classes. The differential expression of 18 genes has been controlled using real-time PCR experiments validating the micro-array analysis. Finally, the integration of the data using Ingenuity Pathways Analysis identified five gene networks providing descriptive elements of the terminal follicular development. Specifically, we observed: (1) the down-expression of ribosomal protein genes, (2) the genes involved in lipid metabolism and (3) the down-expression of cell morphology and ion-binding genes. In conclusion, this study gives new insight into the gene expression during pig terminal follicular growthin vivoand suggested, in particular, a morphological change in pig granulosa cells accompanying terminal follicular growth.

Selection plans in plant and animal breeding are driven by genetic evaluation. Recent developments suggest using massive genetic marker information, known as “genomic selection.” There is little evidence of its performance, though. We empirically compared three strategies for selection: (1) use of pedigree and phenotypic information, (2) use of genomewide markers and phenotypic information, and (3) the combination of both. We analyzed four traits from a heterogeneous mouse population (http://gscan.well.ox.ac.uk/), including 1884 individuals and 10,946 SNP markers. We used linear mixed models, using extensions of association analysis. Cross-validation techniques were used, providing assumption-free estimates of predictive ability. Sampling of validation and training data sets was carried out across and within families, which allows comparing across- and within-family information. Use of genomewide genetic markers increased predictive ability up to 0.22 across families and up to 0.03 within families. The latter is not statistically significant. These values are roughly comparable to increases of up to 0.57 (across family) and 0.14 (within family) in accuracy of prediction of genetic value. In this data set, within-family information was more accurate than across-family information, and populational linkage disequilibrium was not a completely accurate source of information for genetic evaluation. This fact questions some applications of genomic selection.

National audience

A divergent selection experiment based on the homogeneity of birth weight in rabbit was carried out at the INRA experimental farm of Auzeville. The two lines were created by selecting breeding does and bucks from the female strain AGP22 bred at the Grimaud Freres Selection company for environmental variability of birth weight. This involved a new model incorporating a genotypic value for the mean and a genotypic value for the environmental variance. There was a selection response with a significant difference in the within-litter standard deviation of birth weight between the lines after four generations of selection (6.93 g vs. 8.13 g). The mortality of kits at birth and from birth to weaning was lower in the “homogeneous” line than in the “heterogeneous” line (16.6% vs. 18.6% and 17.7% vs. 32.7%, respectively). The litter size at weaning was higher in the “homogeneous” line (7.22 vs. 5.39). The selection on the homogeneity of birth weight had no significant influence on the other traits. It was concluded that selection for reduced environmental variance of birth weight improved young rabbit survival without reducing the mean of birth weight.

International audience; Three different stages of pig antral follicles have been studied in a granulosa-cell transcriptome analysis on nylon microarrays (1152 clones). The data have been generated from seven RNA follicle pools and several technical replicates were made. The objective of this paper was to state the feasibility of a transcriptomic protocol for the study of folliculogenesis in the pig. A statistical analysis was chosen, relying on the linear mixed model (LMM) paradigm. Low variability within technical replicates was hence checked with a LMM. Relevant genes that might be involved in the studied process were then selected. For the most significant genes, statistical methods such as principal component analysis and unsupervised hierarchical clustering were applied to assess their relevance, and a random forest analysis proved their predictive value. The selection of genes was consistent with previous studies and also allowed the identification of new genes whose role in pig folliculogenesis will be further investigated.

Recent biotechnology advances allow for multiple types of omics data, such as transcriptomic, proteomic or metabolomic data sets to be integrated. The problem of feature selection has been addressed several times in the context of classification, but needs to be handled in a specific manner when integrating data. In this study, we focus on the integration of two-block data that are measured on the same samples. Our goal is to combine integration and simultaneous variable selection of the two data sets in a one-step procedure using a Partial Least Squares regression (PLS) variant to facilitate the biologists' interpretation. A novel computational methodology called ;;sparse PLS" is introduced for a predictive analysis to deal with these newly arisen problems. The sparsity of our approach is achieved with a Lasso penalization of the PLS loading vectors when computing the Singular Value Decomposition. Sparse PLS is shown to be effective and biologically meaningful. Comparisons with classical PLS are performed on a simulated data set and on real data sets. On one data set, a thorough biological interpretation of the obtained results is provided. We show that sparse PLS provides a valuable variable selection tool for highly dimensional data sets.

Little is known about the genetic basis of residual feed intake (RFI) variation in fish, since this trait is highly sensitive to environmental influences, and feed intake of individuals is difficult to measure accurately. The purpose of this work was (i) to assess the genetic variability of RFI estimated by an X-ray technique and (ii) to develop predictive criteria for RFI. Two predictive criteria were tested: loss of body weight during feed deprivation and compensatory growth during re-feeding. Ten heterozygous rainbow trout clones were used. Individual intake and body weight were measured three times at threeweek intervals. Then, individual body weight was recorded after two cycles of a three-week feed deprivation followed by a three-week re-feeding. The ratio of the genetic variance to the phenotypic variance was found high to moderate for growth, feed intake, and RFI (VG/VP = 0.63 ± 0.11, 0.29 ± 0.11, 0.29 ± 0.09, respectively). The index that integrates performances achieved during deprivation and re-feeding periods explained 59% of RFI variations. These results provide a basis for further studies on the origin of RFI differences and show that indirect criteria are good candidates for future selective breeding programs.