Exploiting machine learning methods with monthly routine milk recording data and climatic information to predict subclinical mastitis in Italian Mediterranean buffaloes.

Mastitis has detrimental effects on the world's dairy industry, reducing animal health, milk production and quality, as well as income for farmers. In addition, consumers' growing interest in food safety and rational usage of antibiotics highlights the need to develop novel strategies to improve mastitis detection, prevention, and management. In the present study we applied machine learning (ML) analyses to predict presence or absence of subclinical mastitis in Italian Mediterranean buffaloes, exploiting information collected the previous month during routine milk recording procedures, as well as climatic data. The data set included 3,891 records of 1,038 buffaloes from 6 herds located in Basilicata Region (South Italy). Prediction models were developed using 4 different ML algorithms (Generalized Linear Model, Support Vector Machines, Random Forest, and Neural Network) and 2 data set splitting approaches for the creation of the training and test sets (by record or by animal ID number, always with 80% of the data used for model training and the remaining 20% for model testing). Support Vector Machine was the best method to predict high or low somatic cell count at the subsequent test-day record in the validation set, and therefore it was used to estimate the contribution of each feature to the best model. Independently from the data set splitting approach, the most important features were somatic cell score, differential somatic cell count, electrical conductivity, and milk production. Among climatic data, the most informative were temperature and relative humidity. When the data were split by animal ID, an improvement in models' predictive performance on the test set was observed, suggesting this as the most appropriate data splitting approach in data sets with repeated measures to avoid data leakage. According to different metrics, Neural Network was the best method for making predictions on the test set. Our findings confirmed the promising role of ML methods to improve prevention and surveillance of subclinical mastitis, exploiting the large amount of data currently available to identify animals that would possibly have high somatic cell count the subsequent month.

Expression patterns and genetic variation of the ovine skeletal muscle transcriptome of sheep from five Spanish meat breeds.

The goal of the current study is to analyse the gene expression profile of the ovine skeletal muscle as well as to characterize the genetic variation of transcripts expressed in such tissue. This aim has been achieved by sequencing the longissimus dorsi transcriptomes of 50 sheep distributed in five pools representing the Canaria de Pelo, Roja Mallorquina, Gallega, Xisqueta and Ripollesa Spanish autochthonous breeds. Approximately, 363 million reads per pool have been produced and 71.9-82.9% have been successfully mapped to the ovine genome in a paired-end mode (2 × 75 bp). The 200 most expressed muscle transcripts (≈1% of the total transcript count) account for 51% (Canaria de Pelo) to 67% (Gallega) of the total ovine skeletal muscle mRNA expression. These highly expressed genes play key roles in pathways related with striated muscle contraction, gluconeogenesis, glycolysis, citric acid cycle and respiratory electron transport. RNA-Sequencing of muscle transcripts has also revealed that ~72% of the SNPs detected with this approach are shared by at least two pools, and 10% of them segregate in the five pools under analysis. Most of the substitutions detected by RNA-Seq are synonymous or missense and only a minority are predicted to have consequences on protein function.

Romanian wild boars and Mangalitza pigs have a European ancestry and harbour genetic signatures compatible with past population bottlenecks.

We aimed to analyse the genetic diversity of Romanian wild boars and to compare it with that from other wild boar and pig populations from Europe and Asia. Partial sequencing of the mitochondrial encoded cytochrome b (MT-CYB) gene from 36 Romanian wild boars and 36 domestic pigs (Mangalitza, Bazna and Vietnamese breeds) showed that the diversity of Romanian wild boars and Mangalitza pigs is fairly reduced, and that most of the members of these two populations share a common MT-CYB haplotype. Besides, in strong contrast with the Bazna animals, Romanian wild boars and Mangalitza swine did not carry Asian variants at the MT-CYB locus. The autosomal genotyping of 18 Romanian wild boars with the Illumina Porcine SNP60 BeadChip revealed that their genetic background is fundamentally European, even though signs of a potential Near Eastern ancestry (~25%) were detectable at K = 4 (the most significant number of clusters), but not at higher K-values. Admixture analysis also showed that two wild boars are of a hybrid origin, which could be explained by the mating of feral animals with domestic pigs. Finally, a number of Romanian wild boars displayed long runs of homozygosity, an observation that is consistent with the occurrence of past population bottlenecks and the raise of inbreeding possibly due to overhunting or to the outbreak of infectious diseases.

Population structure of eleven Spanish ovine breeds and detection of selective sweeps with BayeScan and hapFLK.

The goals of the current work were to analyse the population structure of 11 Spanish ovine breeds and to detect genomic regions that may have been targeted by selection. A total of 141 individuals were genotyped with the Infinium 50 K Ovine SNP BeadChip (Illumina). We combined this dataset with Spanish ovine data previously reported by the International Sheep Genomics Consortium (N = 229). Multidimensional scaling and Admixture analyses revealed that Canaria de Pelo and, to a lesser extent, Roja Mallorquina, Latxa and Churra are clearly differentiated populations, while the remaining seven breeds (Ojalada, Castellana, Gallega, Xisqueta, Ripollesa, Rasa Aragonesa and Segureña) share a similar genetic background. Performance of a genome scan with BayeScan and hapFLK allowed us identifying three genomic regions that are consistently detected with both methods i.e. Oar3 (150-154 Mb), Oar6 (4-49 Mb) and Oar13 (68-74 Mb). Neighbor-joining trees based on polymorphisms mapping to these three selective sweeps did not show a clustering of breeds according to their predominant productive specialization (except the local tree based on Oar13 SNPs). Such cryptic signatures of selection have been also found in the bovine genome, posing a considerable challenge to understand the biological consequences of artificial selection.

A mitochondrial analysis reveals distinct founder effect signatures in Canarian and Balearic goats.

In the course of human migrations, domestic animals often have been translocated to islands with the aim of assuring food availability. These founder events are expected to leave a genetic footprint that may be recognised nowadays. Herewith, we have examined the mitochondrial diversity of goat populations living in the Canarian and Balearic archipelagos. Median-joining network analysis produced very distinct network topologies for these two populations. Indeed, a majority of Canarian goats shared a single ancestral haplotype that segregated in all sampled islands, suggesting a single founder effect followed by a stepping-stone pattern of diffusion. This haplotype also was present in samples collected from archaeological assemblies at Gran Canaria and Lanzarote, making evident its widespread distribution in ancient times. In stark contrast, goats from Majorca and Ibiza did not share any mitochondrial haplotypes, indicating the occurrence of two independent founder events. Furthermore, in Majorcan goats, we detected the segregation of the mitochondrial G haplogroup that has only been identified in goats from Egypt, Iran and Turkey. This finding suggests the translocation of Asian and/or African goats to Majorca, possibly as a consequence of the Phoenician and Carthaginian colonisations of this island.

East African pigs have a complex Indian, Far Eastern and Western ancestry.

In this study, we have characterized the mitochondrial diversity of 81 swine from Uganda. Median-joining network analysis of D-loop sequences from these individuals and others characterized in previous studies allowed us to determine that Ugandan pigs cluster with populations from the West (Europe/North Africa), Far East and India. In addition, partial sequencing of the Y-chromosome UTY locus in 18 Ugandan domestic pigs revealed the segregation of a single HY1 lineage that has a cosmopolitan distribution. A Western and Far Eastern ancestry for East African pigs had been already reported, but this is the first study demonstrating an additional contribution from the Indian porcine gene pool. This result is consistent with the high frequency of zebuine alleles in cattle from East Africa. The geographic coordinates of East Africa, at the crossroads of many trading routes that, through the ages, linked Europe, Africa and Asia, might explain the rich and complex genetic heritage of livestock native to this area.

Associations between pig adiponectin (ADIPOQ) genotype and serum lipid levels are modulated by age-specific modifiers.

The adiponectin (ADIPOQ) locus is a positional and functional candidate gene for 2 porcine chromosome 13 (SSC13) QTL influencing cholesterol (CHOL) and low-density lipoprotein (LDL) concentrations in 190-d-old pigs. By sequencing 2.37 kb of the pig ADIPOQ cDNA, we have identified 1 c.*1512G>T 3' untranslated region polymorphism that has been genotyped in a Duroc pig commercial population with records for serum lipid levels at 45 and 190 d of age. Statistical analysis of the data have revealed significant associations between the ADIPOQ genotype and CHOL (P=0.0040) and LDL (P=0.0011) concentrations at 190 d but not at 45 d. In family 3, most of the SSC13 QTL effects on LDL levels at 190 d were explained by the ADIPOQ genotype. We also found an association with triglyceride levels at 45 d (P=0.0060) but not at 190 d. Measurement of allelic mRNA imbalance demonstrated that the G and T alleles are expressed at very similar levels in muscle and fat tissues, indicating that the c.*1512G>T polymorphism does not affect transcript abundance. As a whole, results obtained in the current work as well as previous data gathered in humans and pigs provide evidence that the magnitude of associations between blood lipid phenotypes and candidate loci genotypes may vary depending on the age of the individual, therefore suggesting the existence of dynamic genotype×environment interactions changing on a temporal scale.

Technical note: Advantages and limitations of authenticating Palmera goat dairy products by pyrosequencing the melanocortin 1 receptor (MC1R) gene.

Inferring the breed of origin of dairy products can be achieved through molecular analysis of genetic markers with a population-specific pattern of segregation. The goal of the current work was to generate such markers in goats by resequencing several pigmentation genes [melanocortin 1 receptor (MC1R), v-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog (KIT), tyrosinase (TYR), and tyrosinase-related protein 2 (TYRP2)]. This experiment revealed 10 single nucleotide polymorphisms (SNP), including 5 missense mutations and 1 nonsense mutation. These markers were genotyped in 560 goats from 18 breeds originally from Italy, the Iberian Peninsula, the Canary Islands, and North Africa. Although the majority of SNP segregated at moderate frequencies in all populations (including 2 additional markers that were used as a source of information), we identified a c.764G>A SNP in MC1R that displayed highly divergent allelic frequencies in the Palmera breed compared with the Majorera and Tinerfeña breeds from the Canary Islands. Thus, we optimized a pyrosequencing-based technique that allowed us to estimate, very accurately, the allele frequencies of this marker in complex DNA mixtures from different individuals. Once validated, we applied this method to generating breed-specific DNA profiles that made it possible to detect fraudulent cheeses in which Palmero cheese was manufactured with milk from Majorera goats. One limitation of this approach, however, is that it cannot be used to detect illegal manufacturing where Palmero dairy products are produced by mixing milk from Palmera and Majorera goats, because the c.764G>A SNP segregates in both breeds.

Mining the pig genome to investigate the domestication process.

Pig domestication began around 9000 YBP in the Fertile Crescent and Far East, involving marked morphological and genetic changes that occurred in a relatively short window of time. Identifying the alleles that drove the behavioural and physiological transformation of wild boars into pigs through artificial selection constitutes a formidable challenge that can only be faced from an interdisciplinary perspective. Indeed, although basic facts regarding the demography of pig domestication and dispersal have been uncovered, the biological substrate of these processes remains enigmatic. Considerable hope has been placed on new approaches, based on next-generation sequencing, which allow whole-genome variation to be analyzed at the population level. In this review, we provide an outline of the current knowledge on pig domestication by considering both archaeological and genetic data. Moreover, we discuss several potential scenarios of genome evolution under the complex mixture of demography and selection forces at play during domestication. Finally, we highlight several technical and methodological approaches that may represent significant advances in resolving the conundrum of livestock domestication.

Application of the microarray technology to the transcriptional analysis of muscle phenotypes in pigs.

The transcriptome refers to the collection of all transcripts present in a cell. Gene expression has a very dynamic nature; it acts as a bridge between epigenetic marks, DNA sequence and proteins and changes to accommodate the requirements of the cell at each given time. Recent technological advances have created new opportunities to study complex phenotypes from a global point of view. From an animal production perspective, muscle transcriptomics has been investigated in relation to muscle growth, carcass fattening and meat quality traits. In this review, we discuss the impact of nutritional, anatomical and genetic factors on muscle gene expression and meat quality of pigs assessed by microarray technologies. Altogether, several common themes have been revealed by the in-depth analysis of the current body of knowledge, for instance, the involvement of genes related to energy balance and substrate turnover in the oxidative/glycolytic phenotype of red/white muscle fibre types and in the storage of intramuscular fat. The review also covers recent advances in the discovery of expression QTL and regulatory RNAs in porcine breeds as well as technical developments in the field of deep-sequencing technologies that are expected to substantially increase our knowledge about the genetic architecture of meat quality and production traits.

Genetics of serum and muscle lipids in pigs.

Pork meat is one of the most important sources of animal protein in the human diet. Its nutritional properties are partly determined by intramuscular fat content and composition, with existing general consensus about the detrimental effects of cholesterol and saturated fat on cardiovascular health in humans. Because of their physiological resemblance, pigs can be also used as a valuable animal model to study the genetics of human diseases such as atherosclerosis, obesity and dyslipidaemias. Heritability estimates and QTL maps of porcine muscle and serum lipid traits evidence that a considerable amount of genetic variance determining these phenotypes exists, but its molecular basis remains mostly unknown. The recent advent of high-throughput genotyping and sequencing technologies has revolutionised the field of animal genomics. With these powerful tools, finding needles in the genomic haystack has become increasingly feasible. However, these methodological advances should not be deemed as magic bullets. The goal of identifying the many polymorphisms that shape the variability of lipid phenotypes is so challenging that success can be achieved only under the scope of large international consortia.

Genomic architecture of heritability and genetic correlations for intramuscular and back fat contents in Duroc pigs.

Genetic parameters such as heritability and correlations of fat traits in a Duroc population were dissected using molecular markers. The heritabilities of intramuscular fat in 2 muscles, the gluteus medius and LM, and back fat were 0.54, 0.48, and 0.23, respectively. The genetic correlations were well estimated with standardized SNP effects, being 0.65 between intramuscular fat traits and ∼0.37 between any intramuscular fat trait and back fat. Genetic correlations were overestimated when ignoring molecular information. Twelve chromosomes showed additive genetic variance for intramuscular fat compared with 8 for back fat. Population structure was accommodated using 4 different models. The number of significant, P < 5 × 10(-5) (suggestive, P < 2 × 10(-3)), SNP varied across models and ranged from 0 to 4 (2 to 261) for intramuscular fat in the gluteus medius, from 0 to 57 (9 to 564) for intramuscular fat in the LM, and from 3 to 4 (22 to 168) for back fat. Several SNP showed significant deviations from an additive mode of action. Only 2 SNP significantly affected 2 traits simultaneously.