Statistical analysis of individual assignment tests among four cattle breeds using fifteen STR loci1

Detecting SNP markers discriminating horse breeds by deep learning

The assignment of an individual to the true population of origin using a low-panel of discriminant SNP markers is one of the most important applications of genomic data for practical use. The aim of this study was to evaluate the potential of different Artificial Neural Networks (ANNs) approaches consisting Deep Neural Networks (DNN), Garson and Olden methods for feature selection of informative SNP markers from high-throughput genotyping data, that would be able to trace the true breed of unknown samples. The total of 795 animals from 37 breeds, genotyped by using the Illumina SNP 50k Bead chip were used in the current study and principal component analysis (PCA), log-likelihood ratios (LLR) and Neighbor-Joining (NJ) were applied to assess the performance of different assignment methods. The results revealed that the DNN, Garson, and Olden methods are able to assign individuals to true populations with 4270, 4937, and 7999 SNP markers, respectively. The PCA was used to determine how the animals allocated to the groups using all genotyped markers available on 50k Bead chip and the subset of SNP markers identified with different methods. The results indicated that all SNP panels are able to assign individuals into their true breeds. The success percentage of genetic assignment for different methods assessed by different levels of LLR showed that the success rate of 70% in the analysis was obtained by three methods with the number of markers of 110, 208, and 178 tags for DNN, Garson, and Olden methods, respectively. Also the results showed that DNN performed better than other two approaches by achieving 93% accuracy at the most stringent threshold. Finally, the identified SNPs were successfully used in independent out-group breeds consisting 120 individuals from eight breeds and the results indicated that these markers are able to correctly allocate all unknown samples to true population of origin. Furthermore, the NJ tree of allele-sharing distances on the validation dataset showed that the DNN has a high potential for feature selection. In general, the results of this study indicated that the DNN technique represents an efficient strategy for selecting a reduced pool of highly discriminant markers for assigning individuals to the true population of origin.

Comparison of three statistical approaches for feature selection for fine-scale genetic population assignment in four pig breeds

BACKGROUND

Assigning animals to their corresponding breeds through breed informative single-nucleotide polymorphisms (SNPs) is required in many fields. For instance, it is used in the traceability and the authentication of meat and other livestock products. SNPs' information for several pork breeds are now accessible thanks to the availability of dense SNP chips. These SNP chips cover a large number of molecular markers distributed across the entire genome. To identify the pork breed from a sample of industrial meat, one must analyze a large panel of genetic markers depending on the SNP chip used. The analysis of such large datasets requires intensive work. This leads to the idea of creating less dense chips of breed informative markers based on a reduced number of SNPs. Therefore, the analysis of the data emanating from the genotyping of these reduced chips will require less time and effort.

AIM

The objective of this study is to find the most informative SNPs for the discrimination between four pig breeds, namely Duroc, Landrace, Large White, and Pietrain.

METHOD

The Illumina Porcine 60 k SNP chip was used to genotype SNPs distributed all over the individuals' genomes. Firstly, we used three different statistical approaches for feature selection: (i) principal component analysis (PCA), (ii) least absolute shrinkage and selection operator (LASSO), and (iii) random forest (RF). These three approaches identified three sets of SNPs; each set corresponds to one approach. Then, we combined the results of the three methods by setting up a final panel containing the SNPs which appear on the three sets altogether.

RESULTS

Separately, each method resulted in a panel with the corresponding most discriminating SNPs. The PCA, the LASSO, and the random forest with Boruta algorithm highlighted 28,816, 50, and 286 SNPs, respectively. The number of SNPs selected by PCA is high compared to Boruta and LASSO because PCA chooses the variables while preserving as much information about the data as possible. The only downside of LASSO regression is that among a group of correlated variables, LASSO tends to select only one variable and ignore the others regardless of their importance. Contrarily to LASSO, the Boruta algorithm considers the interdependence between SNPs and selects informative variables even if they are correlated and have the same effect. The three panels shared 23 SNPs; the distribution of the individuals according to these SNPs showed a grouping of individuals of each breed in well-defined clusters without any overlapping.

CONCLUSIONS

The biological pathways represented by 23 breed informative SNPs resulted by the combination of PCA, LASSO, and Boruta should be explored in further analysis. The results provided by our study are promising for further applications of this method in other livestock animals.

Detection of selection signatures in Piemontese and Marchigiana cattle, two breeds with similar production aptitudes but different selection histories

BACKGROUND

Domestication and selection are processes that alter the pattern of within- and between-population genetic variability. They can be investigated at the genomic level by tracing the so-called selection signatures. Recently, sequence polymorphisms at the genome-wide level have been investigated in a wide range of animals. A common approach to detect selection signatures is to compare breeds that have been selected for different breeding goals (i.e. dairy and beef cattle). However, genetic variations in different breeds with similar production aptitudes and similar phenotypes can be related to differences in their selection history.

METHODS

In this study, we investigated selection signatures between two Italian beef cattle breeds, Piemontese and Marchigiana, using genotyping data that was obtained with the Illumina BovineSNP50 BeadChip. The comparison was based on the fixation index (Fst), combined with a locally weighted scatterplot smoothing (LOWESS) regression and a control chart approach. In addition, analyses of Fst were carried out to confirm candidate genes. In particular, data were processed using the varLD method, which compares the regional variation of linkage disequilibrium between populations.

RESULTS

Genome scans confirmed the presence of selective sweeps in the genomic regions that harbour candidate genes that are known to affect productive traits in cattle such as DGAT1, ABCG2, CAPN3, MSTN and FTO. In addition, several new putative candidate genes (for example ALAS1, ABCB8, ACADS and SOD1) were detected.

CONCLUSIONS

This study provided evidence on the different selection histories of two cattle breeds and the usefulness of genomic scans to detect selective sweeps even in cattle breeds that are bred for similar production aptitudes.

Genetic diversity, structure and individual assignment of Casta Navarra cattle: a well-differentiated fighting bull population

The Casta Navarra lineage was one of the populations used to establish the fighting bull (FB) breed, and it has also been reproductively isolated from the others FBs. A total of 1284 individuals from two generations of 16 Casta Navarra herds were sampled to analyse their diversity, their genetic structure and the ability of 28 microsatellite markers to assign individuals to closely related populations. These animals were compared with closely related phylogenetic (FB) or geographical (Pirenaica and Monchina) populations. Hardy-Weinberg equilibrium analysis showed that 82% of the loci had a significant heterozygote deficit as a consequence of the Wahlund effect. The average proportion of genetic variation explained by farm differences was 9% by Wright's FST index. A phylogenetic tree constructed with a neighbour-joining method based on Reynolds genetic distances and a Bayesian Markov chain Monte Carlo clustering approach revealed clear differences between farm groups that generally corresponded to historical information and could unambiguously differentiate Casta Navarra cattle from the other populations. The percentage of animals correctly assigned to the Casta Navarra population was 91.78% for a q threshold of >0.9. Admixture was only detected in 4.45% (q < 0.8) of the cattle. These results are relevant for the maintenance and development of diversity and conservation in the Casta Navarra population.

Development of a genetic tool for product regulation in the diverse British pig breed market

Background

The application of DNA markers for the identification of biological samples from both human and non-human species is widespread and includes use in food authentication. In the food industry the financial incentive to substituting the true name of a food product with a higher value alternative is driving food fraud. This applies to British pork products where products derived from traditional pig breeds are of premium value. The objective of this study was to develop a genetic assay for regulatory authentication of traditional pig breed-labelled products in the porcine food industry in the United Kingdom.

Results

The dataset comprised of a comprehensive coverage of breed types present in Britain: 460 individuals from 7 traditional breeds, 5 commercial purebreds, 1 imported European breed and 1 imported Asian breed were genotyped using the PorcineSNP60 beadchip. Following breed-informative SNP selection, assignment power was calculated for increasing SNP panel size. A 96-plex assay created using the most informative SNPs revealed remarkably high genetic differentiation between the British pig breeds, with an average F_ST of 0.54 and Bayesian clustering analysis also indicated that they were distinct homogenous populations. The posterior probability of assignment of any individual of a presumed origin actually originating from that breed given an alternative breed origin was > 99.5% in 174 out of 182 contrasts, at a test value of log(LR) > 0. Validation of the 96-plex assay using independent test samples of known origin was successful; a subsequent survey of market samples revealed a high level of breed label conformity.

Conclusion

The newly created 96-plex assay using selected markers from the PorcineSNP60 beadchip enables powerful assignment of samples to traditional breed origin and can effectively identify mislabelling, providing a highly effective tool for DNA analysis in food forensics.

Identification of mitochondrial markers for genetic traceability of European wild boars and Iberian and Duroc pigs

Iberian pigs and wild boars are the source of highly priced meat and dry-cured products. Iberian maternal origin is mandatory for labeled Iberian products, making necessary the authentication of their maternal breed origin. Discrimination between wild and domestic pig maternal origin may be useful to distinguish labeled wild boar meat obtained from hunting or farming. In order to detect useful polymorphisms to trace Iberian, Duroc and wild boar maternal lineages, we herein investigated the complete porcine mitochondrial DNA (mtDNA) using three complementary approaches. Near-complete mtDNA sequences (16989 bp), excluding the minisatellite present in the displacement loop region (D-loop), were successfully determined in six Iberian pigs, two Duroc and six European wild boars. To complete the mtDNA analysis, the D-loop minisatellite region was also analyzed in the same set of samples by amplification and capillary electrophoresis detection. Finally, the frequencies of Asian and European Cytochrome B (Cyt B) haplotypes were estimated in Iberian (n = 96) and Duroc (n = 125) breeds. Comparison of near-complete mtDNA sequences revealed a total of 57 substitutions and two Indels. Out of them, 32 polymorphisms were potential Iberian markers, 10 potential Duroc markers and 16 potential wild boar markers. Fourteen potential markers (five Iberian and nine Duroc), were selected to be genotyped in 96 Iberian and 91 Duroc samples. Five wild boar potential markers were selected and tested in samples of wild boars (73) and domestic pigs including: 96 Iberian, 16 Duroc, 16 Large White and 16 Landrace. Genotyping results showed three linked markers (m.7998C>T, m.9111T>C, m.14719A>G) absent in Duroc and present in Iberian pigs with a frequency 0.72. Six markers (m.8158C>T, m.8297T>C, m.9230G>A, m.11859A>G, m.13955T>C, m.16933T>C), three of them linked, were absent in Iberian pigs and present in Duroc with a joint frequency of almost 0.50. Finally three linked markers (m.7188G>A, m.9224T>C, m.15823A>G) were solely detected in wild boars with a frequency 0.22. The D-loop minisatellite results showed overlapping ranges of fragment sizes and suggested heteroplasmy, a result that nullify the use of this region for the development of breed diagnostic markers. The Cyt B haplotype results showed the presence of European haplotypes in Iberian while one of the Asian haplotypes was detected in Duroc with a frequency 0.22, linked to the Duroc marker m.9230G>A. Our results are valuable to resolve the problems of Iberian and wild boar maternal origin determination but additional markers are required to achieve totally useful genetic tests.

Assessing priorities for conservation in Tuscan cattle breeds using microsatellites

Preservation of rare genetic stocks requires assessment of within-population genetic diversity and between-population differentiation to make inferences on their degree of uniqueness. A total of 194 Tuscan cattle (44 Calvana, 35 Chianina, 25 Garfagnina, 31 Maremmana, 31 Mucca Pisana and 28 Pontremolese) individuals were genotyped for 34 microsatellite markers. Moreover, 56 samples belonging to Argentinean Creole and Asturiana de la Montaña cattle breeds were used as an outgroup. Genetic diversity was quantified in terms of molecular coancestry and allelic richness. STRUCTURE analyses showed that the Tuscan breeds have well-differentiated genetic backgrounds, except for the Calvana and Chianina breeds, which share the same genetic ancestry. The between-breed Nei's minimum distance (Dm) matrices showed that the pair Calvana-Chianina was less differentiated (0.049 ± 0.006). The endangered Tuscan breeds (Calvana, Garfagnina, Mucca Pisana and Pontremolese) made null or negative contributions to diversity, except for the Mucca Pisana contribution to allelic richness (CT = 1.8%). The Calvana breed made null or negative within-breed contributions (W = 0.0%; CW = -0.4%). The Garfagnina and Pontremolese breeds made positive contributions to between-breed diversity but negative and high within-breed contributions, thus suggesting population bottleneck with allelic losses and increase of homozygosity in the population. Exclusion of the four endangered Tuscan cattle breeds did not result in losses in genetic diversity (T = -0.7%; CT = -1.2%), whereas exclusion of the non-endangered breeds (Chianina and Maremmana) did (T = 2.1%; CT = 3.9%); the simple exclusion of the Calvana breed from the former group led to losses in genetic diversity (T = 0.47%; CT = 2.34%), indicating a diverse significance for this breed. We showed how quantifying both within-population diversity and between-population differentiation in terms of allelic frequencies and allelic richness provides different and complementary information on the genetic backgrounds assessed and may help to implement priorities and strategies for conservation in livestock.

Living together but remaining apart: Atlantic and Mediterranean loggerhead sea turtles (Caretta caretta) in shared feeding grounds

Juvenile loggerhead sea turtles (Caretta caretta) from Atlantic nesting populations migrate into the western Mediterranean, where they share feeding grounds with turtles originating in the Mediterranean. In this scenario, male-mediated gene flow may lead to the homogenization of these distant populations. To test this hypothesis, we genotyped 7 microsatellites from 56 Atlantic individuals sampled from feeding grounds in the western Mediterranean and then compared the observed allele frequencies with published data of 112 individuals from Mediterranean nesting beaches. Mediterranean populations were found to be genetically differentiated from the Atlantic stock reaching the western Mediterranean (F(st) = 0.029, P < 0.001); therefore, the possible mating events between Atlantic and Mediterranean individuals are not sufficient to homogenize these 2 areas. The differentiation observed between these 2 areas demonstrates that microsatellites are sufficiently powerful for mixed stock analysis and that individual assignment (IA) tests can be performed in combination with mitochondrial DNA (mtDNA) analysis. In a set of 197 individuals sampled in western Mediterranean feeding grounds, 87% were robustly assigned to Atlantic or Mediterranean groups with the combined marker, as compared with only 52% with mtDNA alone. These findings provide a new approach for tracking the movements of these oceanic migrants and have strong implications for the conservation of the species.

Evaluation of approaches for identifying population informative markers from high density SNP chips

Background

Genetic markers can be used to identify and verify the origin of individuals. Motivation for the inference of ancestry ranges from conservation genetics to forensic analysis. High density assays featuring Single Nucleotide Polymorphism (SNP) markers can be exploited to create a reduced panel containing the most informative markers for these purposes. The objectives of this study were to evaluate methods of marker selection and determine the minimum number of markers from the BovineSNP50 BeadChip required to verify the origin of individuals in European cattle breeds. Delta, Wright's F_ST, Weir & Cockerham's F_ST and PCA methods for population differentiation were compared. The level of informativeness of each SNP was estimated from the breed specific allele frequencies. Individual assignment analysis was performed using the ranked informative markers. Stringency levels were applied by log-likelihood ratio to assess the confidence of the assignment test.

Results

A 95% assignment success rate for the 384 individually genotyped animals was achieved with < 80, < 100, < 140 and < 200 SNP markers (with increasing stringency threshold levels) across all the examined methods for marker selection. No further gain in power of assignment was achieved by sampling in excess of 200 SNP markers. The marker selection method that required the lowest number of SNP markers to verify the animal's breed origin was Wright's F_ST (60 to 140 SNPs depending on the chosen degree of confidence). Certain breeds required fewer markers (< 100) to achieve 100% assignment success. In contrast, closely related breeds require more markers (~200) to achieve > 95% assignment success. The power of assignment success, and therefore the number of SNP markers required, is dependent on the levels of genetic heterogeneity and pool of samples considered.

Conclusions

While all SNP selection methods produced marker panels capable of breed identification, the power of assignment varied markedly among analysis methods. Thus, with effective exploration of available high density genetic markers, a diagnostic panel of highly informative markers can be produced.

Use of linked loci as individuals or haplotypes for marker-assisted breed assignment

The objective of this study was to use simulation to evaluate the benefits of considering haplotypes of loci when linked single nucleotide polymorphisms are used for breed assignment. Three breeds of 10,000 females each were simulated under eight scenarios that differed according to the number of generations separating the breeds, size of breed founder populations and recombination rate between linked loci. Molecular genotypes consisted of 20 groups of three linked loci each. Breed assignment was performed in the final generation and was based on the frequency method. Haplotypes were reconstructed using the expectation-maximization algorithm. Accuracy of breed assignment was based on the frequency of correct breed assignment. Assignment accuracy increased as more genotypes (loci or haplotypes) were considered and more animals were used to estimate genotypic frequencies within breed. For most scenarios, use of haplotypes yielded equal or greater accuracies than when loci were considered independent. The advantage of haplotypes tended to increase as linkage disequilibrium between adjacent loci increased. The greatest advantage for using haplotypes was observed when recombination rate was low (0.001), breeds were separated by few generations (100), and a relatively large number of founder animals (110) was used to form new breeds. In this situation, 90% accuracy of breed assignment was achieved using nine to 14 haplotypes (i.e. 27-42 loci) depending on breed, vs. 39-57 individual loci.

Genetic traceability of livestock products: A review

Traceability is the ability to maintain the identification of animal, or animal products, all along the production chain. It represents an essential tool to safeguard public and animal health and to valorize typical production systems. European food legislation is particularly strict and traceability systems, based on product labeling, have become mandatory in all European countries. However, the implementation of this system does not ensure consumers against fraud. Paper documents can be counterfeit so researchers have focused on the study of genetic traceability systems based on products identification through DNA analysis. In fact DNA is inalterable, detectable in every cell, resistant to heat treatments, and allows for individual, breed or species identification. Even if results are promising, these techniques are too expensive to be converted in routine tests but they could be a trusted tool for verification of suspected fraud. The present review proposes a synthesis of the major advances made in individual, breed, and species genetic identification in the last years, focusing on advantages and disadvantages and on their real future applications for animal productions.

Breed assignment of Italian cattle using biallelic AFLP markers

The verification of the breed origin of animal products is relevant for food safety and authenticity. We assessed the suitability of AFLP molecular markers in the assignment of cattle individuals to their breed of origin. Three hundred and ninety-six animals belonging to 16 cattle breeds genotyped with 141 AFLP markers were used as reference data set. Assignment was performed with likelihood (aflpop) and Bayesian (structure) methods. The Bayesian approach was superior to the likelihood algorithm with respect to (i) the correct assignment of simulated individuals to their breed of origin (93% vs. 81% respectively), (ii) the correct assignment of 44 sampled Romagnola animals (91% vs. 45% respectively) and (iii) the correct classification of animals belonging to a breed that was not included within the reference dataset. Thus, AFLP profiling in combination with the Bayesian approach seems a useful tool for breed assignment.