Cali E. Willet

Genome-Wide Association Studies in Dogs and Humans Identify ADAMTS20 as a Risk Variant for Cleft Lip and Palate

Cleft lip with or without cleft palate (CL/P) is the most commonly occurring craniofacial birth defect. We provide insight into the genetic etiology of this birth defect by performing genome-wide association studies in two species: dogs and humans. In the dog, a genome-wide association study of 7 CL/P cases and 112 controls from the Nova Scotia Duck Tolling Retriever (NSDTR) breed identified a significantly associated region on canine chromosome 27 (unadjusted p=1.1 x 10-13; adjusted p= 2.2 x 10-3). Further analysis in NSDTR families and additional full sibling cases identified a 1.44 Mb homozygous haplotype (chromosome 27: 9.29 – 10.73 Mb) segregating with a more complex phenotype of cleft lip, cleft palate, and syndactyly (CLPS) in 13 cases. Whole-genome sequencing of 3 CLPS cases and 4 controls at 15X coverage led to the discovery of a frameshift mutation within ADAMTS20 (c.1360_1361delAA (p.Lys453Ilefs*3)), which segregated concordant with the phenotype. In a parallel study in humans, a family-based association analysis (DFAM) of 125 CL/P cases, 420 unaffected relatives, and 392 controls from a Guatemalan cohort, identified a suggestive association (rs10785430; p =2.67 x 10-6) with the same gene, ADAMTS20. Sequencing of cases from the Guatemalan cohort was unable to identify a causative mutation within the coding region of ADAMTS20, but four coding variants were found in additional cases of CL/P. In summary, this study provides genetic evidence for a role of ADAMTS20 in CL/P development in dogs and as a candidate gene for CL/P development in humans.

Development of a SNP-based assay for measuring genetic diversity in the Tasmanian devil insurance population.

BackgroundThe Tasmanian devil (Sarcophilus harrisii) has undergone a recent, drastic population decline due to the highly contagious devil facial tumor disease. The tumor is one of only two naturally occurring transmissible cancers and is almost inevitably fatal. In 2006 a disease-free insurance population was established to ensure that the Tasmanian devil is protected from extinction. The insurance program is dependent upon preserving as much wild genetic diversity as possible to maximize the success of subsequent reintroductions to the wild. Accurate genotypic data is vital to the success of the program to ensure that loss of genetic diversity does not occur in captivity. Until recently, microsatellite markers have been used to study devil population genetics, however as genetic diversity is low in the devil and potentially decreasing in the captive population, a more sensitive genotyping assay is required.MethodsUtilising the devil reference genome and whole genome re-sequencing data, we have identified polymorphic regions for use in a custom genotyping assay. These regions were amplified using PCR and sequenced on the Illumina MiSeq platform to refine a set a markers to genotype the Tasmanian devil insurance population.ResultsWe have developed a set of single nucleotide polymorphic (SNP) markers, assayed by amplicon sequencing, that provide a high-throughput method for monitoring genetic diversity and assessing familial relationships among devils. To date we have used a total of 267 unique SNPs within both putatively neutral and functional loci to genotype 305 individuals in the Tasmanian devil insurance population. We have used these data to assess genetic diversity in the population as well as resolve the parentage of 21 offspring.ConclusionsOur molecular data has been incorporated with studbook management practices to provide more accurate pedigree information and to inform breeding recommendations. The assay will continue to be used to monitor the genetic diversity of the insurance population of Tasmanian devils with the aim of reducing inbreeding and maximizing success of reintroductions to the wild.

Canine disorder mirrors human disease: exonic deletion in HES7 causes autosomal recessive spondylocostal dysostosis in miniature Schnauzer dogs.

Spondylocostal dysostosis is a congenital disorder of the axial skeleton documented in human families from diverse racial backgrounds. The condition is characterised by truncal shortening, extensive hemivertebrae and rib anomalies including malalignment, fusion and reduction in number. Mutations in the Notch signalling pathway genes DLL3, MESP2, LFNG, HES7 and TBX6 have been associated with this defect. In this study, spondylocostal dysostosis in an outbred family of miniature schnauzer dogs is described. Computed tomography demonstrated that the condition mirrors the skeletal defects observed in human cases, but unlike most human cases, the affected dogs were stillborn or died shortly after birth. Through gene mapping and whole genome sequencing, we identified a single-base deletion in the coding region of HES7. The frameshift mutation causes loss of functional domains essential for the oscillatory transcriptional autorepression of HES7 during somitogenesis. A restriction fragment length polymorphism test was applied within the immediate family and supported a highly penetrant autosomal recessive mode of inheritance. The mutation was not observed in wider testing of 117 randomly sampled adult miniature schnauzer and six adult standard schnauzer dogs; providing a significance of association of P raw = 4.759e-36 (genome-wide significant). Despite this apparently low frequency in the Australian population, the allele may be globally distributed based on its presence in two unrelated sires from geographically distant locations. While isolated hemivertebrae have been observed in a small number of other dog breeds, this is the first clinical and genetic diagnosis of spontaneously occurring spondylocostal dysostosis in a non-human mammal and offers an excellent model in which to study this devastating human disorder. The genetic test can be utilized by dog breeders to select away from the disease and avoid unnecessary neonatal losses.

Variants in the host genome may inhibit tumour growth in devil facial tumours: evidence from genome-wide association

Devil facial tumour disease (DFTD) has decimated wild populations of Tasmanian devils (Sarcophilus harrisii) due to its ability to avoid immune detection and pass from host to host by biting. A small number of devils have been observed to spontaneously recover from the disease which is otherwise fatal. We have sequenced the genomes of these rare cases and compared them to the genomes of devils who succumbed to the disease. Genome-wide association, based on this limited sampling, highlighted two key genomic regions potentially associated with ability to survive DFTD. Following targeted genotyping in additional samples, both of these loci remain significantly different between cases and controls, with the PAX3 locus retaining significance at the 0.001 level, though genome-wide significance was not achieved. We propose that PAX3 may be involved in a regulatory pathway that influences the slowing of tumour growth and may allow more time for an immune response to be mounted in animals with regressed tumours. This provides an intriguing hypothesis for further research and could provide a novel route of treatment for this devastating disease.

An updated felCat5 SNP manifest for the Illumina Feline 63k SNP genotyping array

Genotypes and analysis: The deletion variant was g.29500068_29500069delAT relative to NC_019477.1 (rs397514112). Genotyping was performed as described (Table S1). Analysis employed the mixed model or glimmix procedure in SAS v9.2 (SAS Institute) as described. For each analysis, the production trait of interest was included as the dependent variable. Other model terms included breed, sire, year of birth, age in years, age at last lambing and genotype as described (Appendix S1). Bonferroni correction accounted for multiple testing.

Exclusion of known progressive retinal atrophy genes for blindness in the Hungarian Puli

goats (P = 0.002) (Table S1). Considering the primary function of GHR, it is plausible that the heterozygous 14-bp deletion decreases body weight and that this explains the lower litter size. Occurrence of the same deletion in Saanen dairy goats and the Xinjiang cashmere goats (Tables S2 & S3) suggests that our findings are also relevant for other goat breeds. We also found significant difference in genotype distribution between these breeds (P < 0.05) (Table S3). We propose the 14-bp duplicated indel locus as informative molecular markers in marker assisted selection programs for optimal growth and litter size in goats.

Mitochondrial DNA analyses of the saltwater crocodile (Crocodylus porosus) from the Northern Territory of Australia

Abstract. The saltwater crocodile is distributed throughout south-east Asia and Australia. In Australia, it is most abundant in the Northern Territory and Queensland, where it is sustainably farmed for its skins and meat. The aim of this study was to elucidate the relationships and genetic structure among saltwater crocodiles from the Northern Territory of Australia using mitochondrial control region sequences from 61 individuals, representing nine river basins and six of unknown origin, as well as published sequences from other regions. Eight mitochondrial control region haplotypes were identified among both published and novel sequences. Three of the haplotypes appear to be restricted to specimens from northern Australia, with a single haplotype being the most widely dispersed across all river basins. Although Analysis of Molecular Variance provides some support for differentiation among river basins, the frequency of shared haplotypes among these geographical units and median-joining network analysis do not support a clear genetic structure or phylogeographic pattern for saltwater crocodiles in the Northern Territory. The results of this study will assist in furthering our understanding of the genetic diversity of wild saltwater crocodile populations used for ranching in the Northern Territory, as well as providing a framework for assessing the origin of unknown specimens in the future.

From the phenotype to the genotype via bioinformatics.

Moving a project from the status of observing a trait of interest to identifying the underlying causal variant is a challenging task requiring a series of bioinformatics procedures and ideally the availability of a suitable reference genome sequence and its associated resources. We visit common practices for discovering the biology underlying observed traits in mammals.

A Coding Variant in the Gene Bardet-Biedl Syndrome 4 (BBS4) Is Associated with a Novel Form of Canine Progressive Retinal Atrophy

Progressive retinal atrophy is a common cause of blindness in the dog and affects >100 breeds. It is characterized by gradual vision loss that occurs due to the degeneration of photoreceptor cells in the retina. Similar to the human counterpart retinitis pigmentosa, the canine disorder is clinically and genetically heterogeneous and the underlying cause remains unknown for many cases. We use a positional candidate gene approach to identify putative variants in the Hungarian Puli breed using genotyping data of 14 family-based samples (CanineHD BeadChip array, Illumina) and whole-genome sequencing data of two proband and two parental samples (Illumina HiSeq 2000). A single nonsense SNP in exon 2 of BBS4 (c.58A > T, p.Lys20*) was identified following filtering of high quality variants. This allele is highly associated (PCHISQ = 3.425e−14, n = 103) and segregates perfectly with progressive retinal atrophy in the Hungarian Puli. In humans, BBS4 is known to cause Bardet–Biedl syndrome which includes a retinitis pigmentosa phenotype. From the observed coding change we expect that no functional BBS4 can be produced in the affected dogs. We identified canine phenotypes comparable with Bbs4-null mice including obesity and spermatozoa flagella defects. Knockout mice fail to form spermatozoa flagella. In the affected Hungarian Puli spermatozoa flagella are present, however a large proportion of sperm are morphologically abnormal and <5% are motile. This suggests that BBS4 contributes to flagella motility but not formation in the dog. Our results suggest a promising opportunity for studying Bardet–Biedl syndrome in a large animal model.

Simple, rapid and accurate genotyping-by-sequencing from aligned whole genomes with ArrayMaker

SUMMARY Whole-genome sequencing has revolutionized the study of genetics. Genotyping-by-sequencing is now a viable method of genotyping, yet the bioinformatics involved can be daunting if not prohibitive for some laboratories. Here we present ArrayMaker, a user-friendly tool that extracts accurate single nucleotide polymorphism genotypes at pre-defined loci from whole-genome alignments and presents them in a standard genotyping format compatible with association analysis software and datasets genotyped on commercial array platforms. Using this tool, geneticists with only basic computing ability can genotype samples at any desired list of markers, facilitating genome-wide association analysis, fine mapping, candidate variant assessment, data sharing and compatibility of data sourced from multiple technologies. AVAILABILITY AND IMPLEMENTATION ArrayMaker is licensed under The MIT License and can be freely obtained at https://github.com/cw2014/ArrayMaker/. The program is implemented in Perl and runs on Linux operating systems. SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online. CONTACT cali.willet@sydney.edu.au.