Advances in understanding the genetic architecture of antibody response to paratuberculosis in sheep by heritability estimate and LDLA mapping analyses and investigation of candidate regions using sequence-based data.

BACKGROUND: Paratuberculosis is a contagious and incurable disease that is caused by Mycobacterium avium subsp. paratuberculosis (MAP) with significant negative effects on animal welfare and farm profitability. Based on a large naturally infected flock over 12 years, we analyzed repeated enzyme-linked immunosorbent assay tests (ELISA), OvineSNP50 BeadChip genotypes and whole-genome sequences imputed from 56 influential animals. The main goals were to estimate the genetic parameters of proxy traits for resistance to MAP, identify genomic regions associated with the host's immune response against MAP and search for candidate genes and causative mutations through association and functional annotation analyses of polymorphisms identified by sequencing. RESULTS: Two variables were derived from ELISA tests. The first, a binary variable, assessed the infection status of each animal over the entire productive life, while the second considered the level of antibody recorded over time. Very similar results were obtained for both variables. Heritability estimates of about 0.20 were found and a significant region capturing 18% and 13% of the genetic variance was detected on ovine chromosome 20 by linkage disequilibrium and linkage analysis on OvineSNP50 positions. Functional annotation and association analyses on the imputed sequence polymorphisms that were identified in this region were carried out. No significant variants showed a functional effect on the genes that mapped to this region, most of which belong to the major histocompatibility complex class II (MHC II). However, the conditional analysis led to the identification of two significant polymorphisms that can explain the genetic variance associated with the investigated genomic region. CONCLUSIONS: Our results confirm the involvement of the host's genetics in susceptibility to MAP in sheep and suggest that selective breeding may be an option to limit the infection. The estimated heritability is moderate with a relevant portion being due to a highly significant region on ovine chromosome 20. The results of the combined use of sequence-based data and functional analyses suggest several genes belonging to the MHC II as the most likely candidates, although no mutations in their coding regions showed a significant association. Nevertheless, information from genotypes of two highly significant polymorphisms in the region can enhance the efficiency of selective breeding programs.

Association analysis and functional annotation of imputed sequence data within genomic regions influencing resistance to gastro-intestinal parasites detected by an LDLA approach in a nucleus flock of Sarda dairy sheep.

BACKGROUND: Gastroinestinal nematodes (GIN) are one of the major health problem in grazing sheep. Although genetic variability of the resistance to GIN has been documented, traditional selection is hampered by the difficulty of recording phenotypes, usually fecal egg count (FEC). To identify causative mutations or markers in linkage disequilibrium (LD) to be used for selection, the detection of quantitative trait loci (QTL) for FEC based on linkage disequilibrium-linkage analysis (LDLA) was performed on 4097 ewes (from 181 sires) all genotyped with the OvineSNP50 Beadchip. Identified QTL regions (QTLR) were imputed from whole-genome sequences of 56 target animals of the population. An association analysis and a functional annotation of imputed polymorphisms in the identified QTLR were performed to pinpoint functional variants with potential impact on candidate genes identified from ontological classification or differentially expressed in previous studies. RESULTS: After clustering close significant locations, ten QTLR were defined on nine Ovis aries chromosomes (OAR) by LDLA. The ratio between the ANOVA estimators of the QTL variance and the total phenotypic variance ranged from 0.0087 to 0.0176. QTL on OAR4, 12, 19, and 20 were the most significant. The combination of association analysis and functional annotation of sequence data did not highlight any putative causative mutations. None of the most significant SNPs showed a functional effect on genes' transcript. However, in the most significant QTLR, we identified genes that contained polymorphisms with a high or moderate impact, were differentially expressed in previous studies, contributed to enrich the most represented GO process (regulation of immune system process, defense response). Among these, the most likely candidate genes were: TNFRSF1B and SELE on OAR12, IL5RA on OAR19, IL17A, IL17F, TRIM26, TRIM38, TNFRSF21, LOC101118999, VEGFA, and TNF on OAR20. CONCLUSIONS: This study performed on a large experimental population provides a list of candidate genes and polymorphisms which could be used in further validation studies. The expected advancements in the quality of the annotation of the ovine genome and the use of experimental designs based on sequence data and phenotypes from multiple breeds that show different LD extents and gametic phases may help to identify causative mutations.

Mapping genomic regions affecting milk traits in Sarda sheep by using the OvineSNP50 Beadchip and principal components to perform combined linkage and linkage disequilibrium analysis.

BACKGROUND: The detection of regions that affect quantitative traits (QTL), to implement selection assisted by molecular information, remains of particular interest in dairy sheep for which genetic gain is constrained by the high costs of large-scale phenotype and pedigree recording. QTL detection based on the combination of linkage disequilibrium and linkage analysis (LDLA) is the most suitable approach in family-structured populations. The main issue in performing LDLA mapping is the handling of the identity-by-descent (IBD) probability matrix. Here, we propose the use of principal component analysis (PCA) to perform LDLA mapping for milk traits in Sarda dairy sheep. METHODS: A resource population of 3731 ewes belonging to 161 sire families and genotyped with the OvineSNP50 Beadchip was used to map genomic regions that affect five milk traits. The paternally and maternally inherited gametes of genotyped individuals were reconstructed and IBD probabilities between them were defined both at each SNP position and at the genome level. A QTL detection model fitting fixed effects of principal components that summarize IBD probabilities was tested at each SNP position. Genome-wide (GW) significance thresholds were determined by within-trait permutations. RESULTS: PCA resulted in substantial dimensionality reduction, in fact 137 and 32 (on average) principal components were able to capture 99% of the IBD variation at the locus and genome levels, respectively. Overall, 2563 positions exceeded the 0.05 GW significance threshold for at least one trait, which clustered into 75 QTL regions most of which affected more than one trait. The strongest signal was obtained for protein content on Ovis aries (OAR) chromosome 6 and overlapped with the region that harbours the casein gene cluster. Additional interesting positions were identified on OAR4 for fat content and on OAR11 for the three yield traits. CONCLUSIONS: PCA is a good strategy to summarize IBD probabilities. A large number of regions associated to milk traits were identified. The outputs provided by the proposed method are useful for the selection of candidate genes, which need to be further investigated to identify causative mutations or markers in strong LD with them for application in selection programs assisted by molecular information.

Combined approaches to identify genomic regions involved in phenotypic differentiation between low divergent breeds: Application in Sardinian sheep populations.

Selective breeding has led to modifications in the genome of many livestock breeds. In this study, we identified the genomic regions that may explain some of the phenotypic differences between two closely related breeds from Sardinia. A total of 44 animals, 20 Sardinian Ancestral Black (SAB) and 24 Sardinian White (SW), were genotyped using the Illumina Ovine 50K array. A total of 68, 38 and 15 significant markers were identified using the case-control genome-wide association study (GWAS), the Bayesian population differentiation analysis (FST ) and the Rsb metric, respectively. Comparisons among the approaches revealed a total of 22 overlapping markers between GWAS and FST and one marker between GWAS and Rsb. Three markers detected by Rsb were also located near (<2 Mb) to highly significant regions identified by GWAS and FST analyses. Moreover, one candidate marker identified by GWAS and FST approaches was located in a run of homozygosity island that was shared by both breeds. We identified several genes involved in many phenotypic differences (such as stature and growth, reproduction, ear size, coat colour, behaviour) between the two analysed breeds. This study shows that combining several genome-wide approaches could improve discovery of regions involved in the variability of breeding traits and responsible for the phenotypic diversity even between closely related breeds. Overall, the combination of such genome-wide methods can be extended to other livestock breeds that share between them a similar genetic background, to understand the process that shapes the patterns of genetic variability between closely related populations.

A genomic map of climate adaptation in Mediterranean cattle breeds.

Domestic species such as cattle (Bos taurus taurus and B. t. indicus) represent attractive biological models to characterize the genetic basis of short-term evolutionary response to climate pressure induced by their post-domestication history. Here, using newly generated dense SNP genotyping data, we assessed the structuring of genetic diversity of 21 autochtonous cattle breeds from the whole Mediterranean basin and performed genome-wide association analyses with covariables discriminating the different Mediterranean climate subtypes. This provided insights into both the demographic and adaptive histories of Mediterranean cattle. In particular, a detailed functional annotation of genes surrounding variants associated with climate variations highlighted several biological functions involved in Mediterranean climate adaptation such as thermotolerance, UV protection, pathogen resistance or metabolism with strong candidate genes identified (e.g., NDUFB3, FBN1, METTL3, LEF1, ANTXR2 and TCF7). Accordingly, our results suggest that main selective pressures affecting cattle in Mediterranean area may have been related to variation in heat and UV exposure, in food resources availability and in exposure to pathogens, such as anthrax bacteria (Bacillus anthracis). Furthermore, the observed contribution of the three main bovine ancestries (indicine, European and African taurine) in these different populations suggested that adaptation to local climate conditions may have either relied on standing genomic variation of taurine origin, or adaptive introgression from indicine origin, depending on the local breed origins. Taken together, our results highlight the genetic uniqueness of local Mediterranean cattle breeds and strongly support conservation of these populations.

Conservation status and historical relatedness of Italian cattle breeds.

BACKGROUND: In the last 50 years, the diversity of cattle breeds has experienced a severe contraction. However, in spite of the growing diffusion of cosmopolite specialized breeds, several local cattle breeds are still farmed in Italy. Genetic characterization of breeds represents an essential step to guide decisions in the management of farm animal genetic resources. The aim of this work was to provide a high-resolution representation of the genome-wide diversity and population structure of Italian local cattle breeds using a medium-density single nucleotide polymorphism (SNP) array. RESULTS: After quality control filtering, the dataset included 31,013 SNPs for 800 samples from 32 breeds. Our results on the genetic diversity of these breeds agree largely with their recorded history. We observed a low level of genetic diversity, which together with the small size of the effective populations, confirmed that several breeds are threatened with extinction. According to the analysis of runs of homozygosity, evidence of recent inbreeding was strong in some local breeds, such as Garfagnina, Mucca Pisana and Pontremolese. Patterns of genetic differentiation, shared ancestry, admixture events, and the phylogenetic tree, all suggest the presence of gene flow, in particular among breeds that originate from the same geographical area, such as the Sicilian breeds. In spite of the complex admixture events that most Italian cattle breeds have experienced, they have preserved distinctive characteristics and can be clearly discriminated, which is probably due to differences in genetic origin, environment, genetic isolation and inbreeding. CONCLUSIONS: This study is the first exhaustive genome-wide analysis of the diversity of Italian cattle breeds. The results are of significant importance because they will help design and implement conservation strategies. Indeed, efforts to maintain genetic diversity in these breeds are needed. Improvement of systems to record and monitor inbreeding in these breeds may contribute to their in situ conservation and, in view of this, the availability of genomic data is a fundamental resource.

The use of embryonic cells in the treatment of osteochondral defects of the knee: an ovine in vivo study.

PURPOSE: the aim of this study was to determine whether local delivery of embryonic stem-like (ESL) cells into osteochondral defects in the femoral condyles of sheep would enhance regeneration of hyaline articular cartilage. METHODS: male ESL cells embedded in fibrin glue were engrafted into osteochondral defects in the medial condyles (ESL-M) of the left femur in 22 ewes. An identical defect was created in the medial condyle of the contralateral stifle joint and left untreated as a control (empty defect, ED). The ewes were divided into 5 groups. Four sheep each were euthanized at 1, 2, 6, and 12 months from surgery, and 6 ewes were euthanized 24 months post-implantation. To study the effect of varying loads on the long-term regeneration process, an identical defect was also created and ESL cell engraftment performed in the lateral condyle (ESL-L) of the left stifle joint of the animals in the 12- and 24-month groups. The evaluation of regenerated tissue was performed by biomechanical, macroscopic, histological, immunohistochemical (collagen type II) and fluorescent in situ hybridization (FISH) assays. RESULTS: no significant differences were found between treated and control sites in the biomechanical assays at any time point. ESL cell grafts showed significantly greater macroscopic evidence of regeneration as compared to controls at 24 months after surgery; significantly better histological evidence of repair in ESL-M samples versus controls was found throughout the considered period. At 24 months from surgery there was significantly improved integration of graft edges with the host tissue in the ESL-M as compared to the ESL-L samples, demonstrating that load bearing positively affects the long-term regeneration process. CONCLUSIONS: ESL cells enhanced the regeneration of hyaline cartilage. FISH confirmed that the regenerative tissue originated from ESL cells. CLINICAL RELEVANCE: ESL cells are able to self-renew for prolonged periods without differentiation and, most importantly, to differentiate into a large variety of tissues.

Treatment with embryonic stem-like cells into osteochondral defects in sheep femoral condyles.

BACKGROUND: Articular cartilage has poor intrinsic capacity for regeneration because of its avascularity and very slow cellular turnover. Defects deriving from trauma or joint disease tend to be repaired with fibrocartilage rather than hyaline cartilage. Consequent degenerative processes are related to the width and depth of the defect. Since mesenchymal stem cells (MSCs) deriving from patients affected by osteoarthritis have a lower proliferative and chondrogenic activity, the systemic or local delivery of heterologous cells may enhance regeneration or inhibit the progressive loss of joint tissue. Embryonic stem cells (ESCs) are very promising, since they can self-renew for prolonged periods without differentiation and can differentiate into tissues from all the 3 germ layers. To date only a few experiments have used ESCs for the study of the cartilage regeneration in animal models and most of them used laboratory animals. Sheep, due to their anatomical, physiological and immunological similarity to humans, represent a valid model for translational studies. This experiment aimed to evaluate if the local delivery of male sheep embryonic stem-like (ES-like) cells into osteochondral defects in the femoral condyles of adult sheep can enhance the regeneration of articular cartilage. Twenty-two ewes were divided into 5 groups (1, 2, 6, 12 and 24 months after surgery). Newly formed tissue was evaluated by macroscopic, histological, immunohistochemical (collagen type II) and fluorescent in situ hybridization (FISH) assays. RESULTS: Regenerated tissue was ultimately evaluated on 17 sheep. Samples engrafted with ES-like cells had significantly better histologic evidence of regeneration with respect to empty defects, used as controls, at all time periods. CONCLUSIONS: Histological assessments demonstrated that the local delivery of ES-like cells into osteochondral defects in sheep femoral condyles enhances the regeneration of the articular hyaline cartilage, without signs of immune rejection or teratoma for 24 months after engraftment.

XVI(th) QTLMAS: simulated dataset and comparative analysis of submitted results for QTL mapping and genomic evaluation.

BACKGROUND: A common dataset was simulated and made available to participants of the XVI(th) QTL-MAS workshop. Tasks for the participants were to detect QTLs affecting three traits, to assess their possible pleiotropic effects, and to evaluate the breeding values in a candidate population without phenotypes using genomic information. METHODS: Four generations consisting of 20 males and 1000 females were generated by mating each male with 50 females. The genome consisted of 5 chromosomes, each of 100 Mb size and carrying 2,000 equally distributed SNPs. Three traits were simulated in order to mimic milk yield, fat yield and fat content. Genetic (co)variances were generated from 50 QTLs with pleiotropic effects. Phenotypes for all traits were expressed only in females, and were provided for the first 3 generations. Fourteen methods for detecting single-trait QTL and 3 methods for investigating their pleiotropic nature were proposed. QTL mapping results were compared according to the following criteria: number of true QTL detected; number of false positives; and the proportion of the true genetic variance explained by submitted positions. Eleven methods for estimating direct genomic values of the candidate population were proposed. Accuracies and bias of predictions were assessed by comparing estimated direct genomic values with true breeding values. RESULTS: The number of true detections ranged from 0 to 8 across methods and traits, false positives from 0 to 15, and the proportion of genetic variance captured from 0 to 0.82, respectively. The accuracy and bias of genomic predictions varied from 0.74 to 0.85 and from 0.86 to 1.34 across traits and methods, respectively. CONCLUSIONS: The best results in terms of detection power were obtained by ridge regression that, however, led to a large number of false positives. Good results both in terms of true detections and false positives were obtained by the approaches that fit polygenic effects in the model. The investigation of the pleiotropic nature of the QTL permitted the identification of few additional markers compared to the single-trait analyses. Bayesian and grouped regularized regression methods performed similarly for genomic prediction while GBLUP produced the poorest results.

Cryobanking of farm animal gametes and embryos as a means of conserving livestock genetics.

In the last few decades, farm animal genetic diversity has rapidly declined, mainly due to changing market demands and intensification of agriculture. But, since the removal of single species can affect the functioning of global ecosystems, it is in the interest of international community to conserve the livestock genetics and to maintain biodiversity. Increasing awareness on the reduction of breed diversity has prompted global efforts for conservation of farm animal breeds. The goals of conservation are to keep genetic variation as gene combinations in a reversible form and to keep specific genes of interest. For this purpose two types of strategies are usually proposed: in situ and ex situ conservation. In situ conservation is the breed maintaining within the livestock production system, in its environment through the enhancement of its production characteristics. Ex situ in vivo conservation is the safeguard of live animals in zoos, wildlife parks, experimental farms or other specialized centres. Ex situ in vitro conservation is the preservation of genetic material in haploid form (semen and oocytes), diploid (embryos) or DNA sequences. In the last few years, ex situ in vitro conservation programs of livestock genetic resources have focused interest on cryopreservation of gametes, embryos and somatic cells as well as testis and ovarian tissues, effectively lengthening the genetic lifespan of individuals in a breeding program even after the death. However, although significant progress has been made in semen, oocytes and embryo cryopreservation of several domestic species, a standardized procedure has not been established yet. The aim of the present review is to describe the cryobanking purposes, the collection goals, the type of genetic material to store and the reproductive biotechnologies utilized for the cryopreservation of farm animal gametes and embryos.

Alternative strategies for selecting subsets of predicting SNPs by LASSO-LARS procedure.

BACKGROUND: The least absolute shrinkage and selection operator (LASSO) can be used to predict SNP effects. This operator has the desirable feature of including in the model only a subset of explanatory SNPs, which can be useful both in QTL detection and GWS studies. LASSO solutions can be obtained by the least angle regression (LARS) algorithm. The big issue with this procedure is to define the best constraint (t), i.e. the upper bound of the sum of absolute value of the SNP effects which roughly corresponds to the number of SNPs to be selected. Usai et al. (2009) dealt with this problem by a cross-validation approach and defined t as the average number of selected SNPs overall replications. Nevertheless, in small size populations, such estimator could give underestimated values of t. Here we propose two alternative ways to define t and compared them with the "classical" one. METHODS: The first (strategy 1), was based on 1,000 cross-validations carried out by randomly splitting the reference population (2,000 individuals with performance) into two halves. The value of t was the number of SNPs which occurred in more than 5% of replications. The second (strategy 2), which did not use cross-validations, was based on the minimization of the Cp-type selection criterion which depends on the number of selected SNPs and the expected residual variance. RESULTS: The size of the subset of selected SNPs was 46, 189 and 64 for the classical approach, strategy 1 and 2 respectively. Classical and strategy 2 gave similar results and indicated quite clearly the regions were QTL with additive effects were located. Strategy 1 confirmed such regions and added further positions which gave a less clear scenario. Correlation between GEBVs estimated with the three strategies and TBVs in progenies without phenotypes were 0.9237, 0.9000 and 0.9240 for classical, strategy 1 and 2 respectively. CONCLUSIONS: This suggests that the Cp-type selection criterion is a valid alternative to the cross-validations to define the best constraint for selecting subsets of predicting SNPs by LASSO-LARS procedure.

Invited review: Current state of genetic improvement in dairy sheep.

Dairy sheep have been farmed traditionally in the Mediterranean basin in southern Europe, central Europe, eastern Europe, and in Near East countries. Currently, dairy sheep farming systems vary from extensive to intensive according to the economic relevance of the production chain and the specific environment and breed. Modern breeding programs were conceived in the 1960s. The most efficient selection scheme for local dairy sheep breeds is based on pyramidal management of the population with the breeders of nucleus flocks at the top, where pedigree and official milk recording, artificial insemination, controlled natural mating, and breeding value estimation are carried out to generate genetic progress. The genetic progress is then transferred to the commercial flocks through artificial insemination or natural-mating rams. Increasing milk yield is still the most profitable breeding objective for several breeds. Almost all milk is used for cheese production and, consequently, milk content traits are very important. Moreover, other traits are gaining interest for selection: machine milking ability and udder morphology, resistance to diseases (mastitis, internal parasites, scrapie), and traits related to the nutritional value of milk (fatty acid composition). Current breeding programs based on the traditional quantitative approach have achieved appreciable genetic gains for milk yield. In many cases, further selection goals such as milk composition, udder morphology, somatic cell count, and scrapie resistance have been implemented. However, the possibility of including other traits of selective interest is limited by high recording costs. Also, the organizational effort needed to apply the traditional quantitative approach limits the diffusion of current selection programs outside the European Mediterranean area. In this context, the application of selection schemes assisted by molecular information, to improve either traditional dairy traits or traits costly to record, seems to be attractive in dairy sheep. At the moment, the most effective strategy seems to be the strengthening of research projects aimed at finding causal mutations along the genes affecting traits of economic importance. However, genome-wide selection seems to be unfeasible in most dairy sheep breeds.

Sheep embryonic stem-like cells transplanted in full-thickness cartilage defects.

Articular cartilage regeneration is limited. Embryonic stem (ES) cell lines provide a source of totipotent cells for regenerating cartilage. Anatomical, biomechanical, physiological and immunological similarities between humans and sheep make this animal an optimal experimental model. This study examines the repair process of articular cartilage in sheep after transplantation of ES-like cells isolated from inner cell masses (ICMs) derived from in vitro-produced (IVP) vitrified embryos. Thirty-five ES-like colonies from 40 IVP embryos, positive for stage-specific embryonic antigens (SSEAs), were pooled in groups of two or three, embedded in fibrin glue and transplanted into osteochondral defects in the medial femoral condyles of 14 ewes. Empty defect (ED) and cell-free glue (G) in the controlateral stifle joint served as controls. The Y gene sequence was used to detect ES-like cells in the repair tissue by in situ hybridization (ISH). Two ewes were euthanized at 1 month post-operatively, three each at 2 and 6 months and four at 12 months. Repairing tissue was examined by biomechanical, macroscopic, histological, immunohistochemical (collagen type II) and ISH assays. Scores of all treatments showed no statistical significant differences among treatment groups at a given time period, although ES-like grafts showed a tendency toward a better healing process. ISH was positive in all ES-like specimens. This study demonstrates that ES-like cells transplanted into cartilage defects stimulate the repair process to promote better organization and tissue bulk. However, the small number of cells applied and the short interval between surgery and euthanasia might have negatively affected the results.

Seroprevalence, clinical incidence, and molecular and epidemiological characterisation of small ruminant lentivirus in the indigenous Passirian goat in northern Italy.

Eight dairy flocks, comprising a total of 323 indigenous Passirian goats from northern Italy, were examined to determine the seroprevalence and clinical incidence of small ruminant lentivirus (SRLV) infections and to identify the SRLV subtypes. The seroprevalence was 81.5% (55-95%). The clinical incidence was 2.5% (0-8.3%) and was apparently low due to the practice of culling clinically affected animals. Phylogenetic analysis of eight PCR fragments (one sample from each flock) revealed that all proviruses belonged to the SRLV subtype B1, which suggests a common source of infection. Subtype B1 being the only circulating SRLV, coupled with the fact that mixed herd systems are very rare in South Tyrol, gives hope that an eradication programme in goats can be successful even without including sheep as long as sheep are kept strictly and permanently isolated.

Identifying native animals in crossbred populations: the case of the Sardinian goat population.

The aim of this work was to develop a strategy for using a genetic analysis for identifying native animals in regions where local breeds have been crossed with improved breeds and then compare that strategy to the overall morphology and breeding histories of the herds for identifying these animals. The experiment included the Sardinian goat population, which is a crossbred of native animals with the Maltese breed. Whole herds were assigned to Maltese (five herds; 49 animals), crossbred (18 herds; 117 animals) or Sardinian (12 herds; 164 animals) groups. For the genetic analysis, genotypes of 22 microsatellites were determined on 330 animals, and basic measurements of genetic diversity were calculated. Genetic variability in the microsatellites was different in the three groups. High positive F(IS) showed that inbreeding existed in the subpopulations. The index of genetic differentiation, Nei's standard genetic distance and Reynolds' genetic distance were calculated and found to be significantly different between the three groups. The Sardinian and Maltese groups were the most distant whereas the crossbred group was closer to the Sardinian group. The proportion of the genome derived from two ancestral populations (native Sardinian and Maltese) was assessed using the structure software. Animals were assigned to three clusters on the basis of native Sardinian thresholds. A good correspondence between the empirical (morphology and breeding histories) and the objective genetic analysis was found. Both approaches indicate the presence of three different subpopulations in the Sardinian goat population.

Feasibility of a linear scoring method of udder morphology for the selection scheme of Sardinian sheep.

Breeders are increasingly interested in improving the machine milkability of Sardinian dairy sheep by selection for udder morphology. Nine-point linear scales were developed to appraise teat placement, degree of suspension of the udder, udder depth, and degree of separation of the 2 halves. Repeatabilities within and across lactation were estimated on an experimental flock. Ewes were scored at least 3 times a year from 1999 to 2003. Within-lactation repeatabilities were greater than 0.70 for all linear traits, whereas across-lactation repeatabilities ranged from 0.59 to 0.66. The first linear combination of basic traits, obtained by principal component analysis, was highly correlated with teat placement, degree of suspension of the udder, and udder depth and showed an across-lactation repeatability of 0.76. These results indicate that an accurate evaluation of animals is possible by a single, early lifetime score. Genetic parameters of linear udder traits were estimated using a REML method applied to a sire model. Data were selected from 76,984 scores of Sardinian yearling ewes collected by 31 classifiers from 1999 to 2004. Two models were compared, one fitting the contemporary group effect as fixed and the other as random. Heritabilities of udder traits ranged from 0.19 to 0.31. The model with the random contemporary group effect produced slightly higher heritabilities and higher correlations between the sires' estimated breeding values and the daughters' average scores. As a whole, the genetic correlations between udder traits were favorable, indicating that selection for one trait will produce a positive evolution of the overall udder conformation. In particular, the degree of suspension of the udder was highly correlated with udder depth (0.82). Genetic correlations with milk yield were unfavorable but generally low, with the exception of udder depth (-0.48). Genetic trends were estimated using an animal model. Only udder depth showed a negative constant genetic trend. Overall results indicated that genetic improvement of the udder morphology of Sardinian ewes is feasible, with major emphasis on teat placement and degree of udder suspension, traits showing the highest heritabilities and low unfavorable genetic correlations with milk yield.