WO2015033133A1 - Prognostic gene - Google Patents

Abstract

The invention relates to the use of the FAM161Agene as a biomarker for the prognosis of a canine mammal developing progressive retinal atrophy. The invention also relates to in vitro methods of prognosing progressive retinal atrophy in a canine mammal by detecting a genetic variation within the FAM161Agene and to oligonucleotide probes and prognostic kits for use in said method.

Description

PROGNOSTIC GENE

FIELD OF THE INVENTION

The invention relates to the use of the FAM161A gene as a biomarker for the prognosis of a canine mammal developing progressive retinal atrophy. The invention also relates to in vitro methods of prognosing progressive retinal atrophy in a canine mammal by detecting a genetic variation within the FAM161A gene and to oligonucleotide probes and prognostic kits for use in said method.

BACKGROUND OF THE INVENTION

Progressive retinal atrophy (PRA) in animals is the term used for a group of inherited and progressive retinal diseases characterised by progressive retinal degeneration resulting in loss of vision. Typically, rod photoreceptor responses are lost first followed by cone photoreceptor responses [1] . Bilateral and symmetrical fundus changes are observed, including tapetal hyper-reflectivity in the early stages followed by vascular attenuation, pigmentary changes and atrophy of the optic nerve head in the later stages of disease [2] . Numerous forms of PRA have been documented in more than 100 dog breeds and while they exhibit similar clinical signs, the aetiology, age of onset and rate of progression vary between and within breeds. While several disease-causing genes have been reported for some forms of PRA [3], many remain undefined.

Retinitis Pigmentosa (RP), the human equivalent of PRA, is the collective name for a group of inherited human retinal disorders that leads to progressive loss of vision in approximately 1 in 4000 people [4,5,6] . As in PRA, rod photoreceptor cells are predominantly affected resulting in clinical symptoms typically including night blindness and loss of peripheral vision. However, the cones also degenerate eventually, resulting in central vision loss and eventually complete blindness is possible. To date, 192 genes have been shown to cause a wide spectrum of retinal disease, including RP (RetNet; http://www.sph.uth.tmc.edu/retnet/), although mutations in these genes currently only account for approximately 30% of recessive RP cases. [7] . Canine diseases are valuable natural models for the study of many varied human conditions such as cardiac autosomal recessive congenital ichthyosis [8], myotubular myopathy [9] and hereditary retinopathies such as Leber congenital amaurosis (LCA) and achromatopsia [10,11] . Further to this, canine models for human eye diseases have proved invaluable in gene-therapy studies, most notably the canine models of LCA associated with RPE65 [12,13,14,15,16] and X-linked RP associated with Retinitis pigmentosa GTPase regulator (RPGR) [17] .

Most PRA cases in the Tibetan Spaniel (TS) are clinically indistinguishable from other forms of PRA. The mode of inheritance appears from pedigree information to be autosomal recessive and the age of diagnosis is most commonly at a relatively late age of approximately 5 years. No mutations have previously been associated with PRA in the breed.

There is therefore a great need to identify the causal genetic variant responsible for PRA in canines.

SUMMARY OF THE INVENTION

According to a first aspect of the invention, there is provided a use of the FAM161A gene as a biomarker for the in vitro prognosis of progressive retinal atrophy developing in a canine mammal.

According to a further aspect of the invention, there is provided an in vitro method of prognosing progressive retinal atrophy in a canine mammal, the method comprising the step of detecting genetic variation within the FAM161A gene.

According to a further aspect of the invention, there is provided oligonucleotide probes for use in a method of prognosing progressive retinal atrophy in a canine mammal, wherein said oligonucleotide probes detect a genetic variation within the FAM161A gene and are as defined herein. According to a further aspect of the invention, there is provided a kit for use in a method of prognosing progressive retinal atrophy in a canine mammal, wherein said kit comprises:

(a) oligonucleotide probes which detect a genetic variation in the FAM161A gene and are as defined herein; and

(b) means for providing a test sample from the canine mammal.

According to a further aspect of the invention, there is provided a method of treating progressive retinal atrophy in a canine mammal, which comprises assessing the progressive retinal atrophy status of a canine mammal by use of a method as defined herein and if the canine mammal is identified as affected by progressive retinal atrophy, treating said canine mammal to prevent or reduce the onset of progressive retinal atrophy. According to a further aspect of the invention, there is provided a method of treating progressive retinal atrophy in a canine mammal, which comprises increasing the level of non-mutant, wild-type FAM161A gene expression and/or FAM161A gene product activity in the canine mammal.

BRIEF DESCRIPTION OF THE FIGURES

FIGURE 1: Genome-wide association mapping of PRA in Tibetan Spaniels

-LoglO of p-values after correction for population stratification. The dotted lines indicate the Bonferroni-corrected 5% significance level based on 15,674 SNPs. A) The CMH meta-analysis approach shows the strongest signal on CFA10 (p_raw = 2.01xl0^"5, Pgenome = 0.014). B) The FMM approach also shows a prominent signal on CFA10 (p_raw = 5.67xl0^"5). C) The signal spans a region of 5.37 Mb from 62 to 67.37 Mb on CFA10.

FIGURE 2: Critical region definition using homozygosity analysis

Microsatellite and SNP marker genotypes for 18 PRA cases and five PRA controls (including one obligate carrier) over the region identified during the GWA study. The most prevalent allele for each marker is white, while alternate alleles are shaded in black. It is not possible to define a region for which all of the cases are homozygous, however, it is possible to define a broad region for which most of the cases (12/18) are homozygous, from 63.935 to 67.729 Mb. The most associated SNP markers are indicated with arrows (—►).

FIGURE 3: IGV display of the SINE insertion in F AM 161 A

A) Each of the three samples (PRA-affected, obligate carrier and control) viewed in IGV are represented by two panels. The increased read depth (upper panels) in the affected panel is characteristic of a duplication, caused by the repeat motif flanking the insertion. The sudden termination of reads (lower panels) and the insertion symbol (I) either side of the duplicated sequence is also characteristic of an insertion flanked by the duplicated sequence. The inserted sequence is present in all reads in the PRA-affected sample, approximately half the reads in the obligate carrier and none of the reads in the PRA- unaffected (control) sample. B) Inserted sequence (italics) as determined from NGS data, flanked by 14 bp repeats (underlined).

FIGURE 4: Sequence of the SINE insertion

A) The sequence and B) graphical representation of the SINE insertion. The precise number of nucleotides that comprise the underlined portion of the poly(A) tract (bold) remains unclear, but is approximately 35-50.

FIGURE 5: Graphical comparison of the intron-exon boundaries of F AM 161 A

A) Mouse (Mus muscularis) FAM161A. B) Human (Homo sapiens) FAM161A. C) Canine (Cam^'s familiaris) FAM161A as predicted by Ensembl genebuild. Six genebuild exons (black) are identical to the mouse and human exons, while the intron-exon boundaries for five exons (checkered) are inconsistent and four exons (white) bear little or no resemblance to human and mouse exons. D) Canine FAM161A exons confirmed by sequencing the retinal mRNA transcript. Exon 1 (diagonal lines) was only partially sequenced. The location of the splice site insertion is indicated. E) Only the 3'-end of exon 1 (not underlined) was successfully sequenced. Two alternative alleles were identified in retinal mRNA that differed from the CanFam2.0 reference sequence.

FIGURE 6: Comparison of FAM161A mRNA isoforms PCR and electrophoresis to compare the FAM161A isoforms in blood from an affected dog and unaffected dog. A) Primers in exons 3 and 6 (arrows) were used to amplify FAM161A isoforms created by alternative splicing of exons 4 and 5, resulting in four possible amplicons expected (fl, fl-5, sh, sh-5; sizes indicated). B) Agarose gel electrophoresis of PCR amplicons.

FIGURE 7: FAM161A_c.₁₇₅₈-i5_i758-i6\ns238 effect on pre-mRNA splicing

Wild-type and FAM161A^'1' sequence, with exon 5 (bold italics), intron 4 (black), the SINE-flanking repeat (single underlined) and SINE insertion c.1758- 15_1758-16ins238 (bold) indicated. In the wild-type sequence, the putative BPS (double underlined, with the critical A bold) is 76 bp upstream of the intron-exon boundary, while in the affected sequence it is more than 300 bp upstream of the boundary. DETAILED DESCRIPTION OF THE INVENTION

According to a first aspect of the invention, there is provided a use of the FAM161A gene as a biomarker for the in vitro prognosis of progressive retinal atrophy developing in a canine mammal. According to a further aspect of the invention, there is provided an in vitro method of prognosing progressive retinal atrophy in a canine mammal, the method comprising the step of detecting genetic variation within the FAM161A gene. References herein to 'progressive retinal atrophy' may also be referred to as 'PRA' or 'PRAS'. PRA is characterised by progressive retinal degeneration resulting in loss of vision.

The examples provided herein report the identification of a short interspersed nuclear element (SINE) insertion in a ciliary gene known as FAM161A (Family With Sequence Similarity 161, Member A). The mutation causes exon skipping and a subsequent shift in the reading frame resulting in a premature termination codon. The evidence shows that this mutation represents a major susceptibility locus for late onset PRA, known hereafter as PRA3, in Tibetan Spaniels and Tibetan Terriers.

FAM161A has been shown to occur in two main isoforms, full-length (FAM161Au) and short (FAM161A_Sh), formed by alternative splicing of exon 4 (Figure 5). Sequencing revealed that canine FAM161Au contains 716 amino acids (Genbank Accession No. KF177335) and FAM161A_Sh contains 660 amino acids (Genbank Accession No. KF177336), with predicted molecular weights of 83kDa and 76kDa, respectively.

The present inventors have found that the insertion mutation results in a further 2 isoforms which lack exon 5, i.e. FAM 161A_fi_₅ and FAM 161A_sh_₅. These isoforms have been surprisingly found to be associated with dogs affected with PRA. The inventors have developed a genotyping-based prognostic test that can be used to determine whether a dog is clear, will potentially be affected by, or a carrier of progressive retinal atrophy. This can be used, inter alia, in selective breeding to avoid affected offspring. It will be understood that references herein to a "prognostic test" refer to a test that can be used to predict the relative likelihood of a canine mammal developing a particular disease. Progressive retinal atrophy generally has a late onset in canines, therefore the newly identified mutation disclosed herein, has the advantage of being able to predict how likely a young canine is to develop the disease within its lifetime.

Very little is known about the structure and function of FAM161A in visual and other pathways. Further investigations are required to elucidate the precise pathways in which FAM 161A is involved. To this end, the canine model described herein could be particularly useful, as no other animal models have been reported.

It will also be appreciated that the present invention can equally be used to diagnose a canine mammal affected by progressive retinal atrophy. Furthermore, the inventors have devised a prognostic genotyping assay that determines the presence or absence of mutation in the canine FAM161A gene in canine DNA.

In one embodiment, the canine mammal is a dog which is a breed selected from Tibetan Spaniel, Tibetan Terrier and Lhasa Apso. In a further embodiment, the canine mammal is a dog which is a breed selected from Tibetan Spaniel and Tibetan Terrier. In a yet further embodiment, the canine mammal is a Tibetan Spaniel.

In one embodiment, the method comprises the steps of:

(i) providing a sample of nucleic acid from the canine mammal;

(ii) detecting genetic variation within the FAM161A gene in the nucleic acid sample; and

(iii) correlating the result from step (ii) with the prognosis of progressive retinal atrophy developing in the canine mammal .

Particular methods of detecting markers in nucleic acid samples are described in more detail hereinafter.

Nucleic acid sample

The sample from the canine mammal may be prepared from any convenient sample, for example from blood or skin tissue. In one embodiment, DNA is extracted from blood, retinal tissue or from buccal (cheek) cells on a swab. In a further embodiment, DNA is extracted from blood or buccal cells on a swab.

The DNA sample analysed may be all or part of the sample being obtained. Methods of the present invention may therefore include obtaining a sample of nucleic acid obtained from the canine mammal. Alternatively, the assessment of the FAM161A gene may be performed or based on an historical DNA sample, or information already obtained therefrom e.g. by assessing the FAM161A gene in DNA sequences which are stored on a databank. In one embodiment, the nucleic acid is genomic DNA (gDNA). In an alternative embodiment, the nucleic acid is messenger RNA (mRNA). In a further alternative embodiment, the nucleic acid is complementary DNA (cDNA). Genetic Variations

It will be appreciated that the genetic variations include any variation in the native, non-mutant or wild type genetic code of the FAM161A gene from said canine mammal under analysis. Examples of such genetic variations include: mutations {e.g. point mutations), substitutions, deletions, insertions, single nucleotide polymorphisms (SNPs), haplotypes, chromosome abnormalities, Copy Number Variation (CNV), epigenetics and DNA inversions.

In one embodiment the genetic variation is a functional mutation i.e. one which is causative of progressive retinal atrophy. Mutations may be functional in that they affect amino acid encoding, or by disruption of regulatory elements {e.g., which may regulate gene expression, or by disruption of sequences - which may be exonic or intronic - involved in regulation of splicing). However it will be appreciated that other markers showing association with progressive retinal atrophy, may also have diagnostic utility and could be used in combination with the assessment of the invention.

In one embodiment, the genetic variation is an insertion mutation which causes a frameshift in the FAM161A gene. This may cause premature termination and/or result in a truncated protein.

In one embodiment, the genetic variation is within intron 4 of the FAM161A gene. Intron 4 of the FAM161A gene is located from position 64,974,116 to position 64,975,764 on chromosome 10 (CanFam 2.0). In one embodiment, the genetic variation comprises an insertion mutation within the FAM161A gene. In a further embodiment, the insertion occurs near the acceptor splice site of intron 4, i.e. near the boundary of intron 4 and exon 5. In one embodiment, the insertion mutation comprises a SINE insertion. The SINE insertion identified in the Examples described herein was found between position 64,974,130 and 64,974,131 on chromosome 10 (CanFam 2.0), therefore in a further embodiment, the insertion mutation comprises a SINE insertion at position 64,974,130 on chromosome 10.

References herein to a 'SINE insertion' refer to a 'short interspersed nuclear element' which is inserted in the target gene. SINE insertions are usually less than 500 nucleotides long, such as less than 400, 300, 200 or 100 nucleotides long. SINE insertions are usually followed by a dinucleotide repeat and a poly(A) tract at their target site.

In one embodiment, detection of the genetic variation is performed by detecting the absence of exon 5 in mRNA of the FAM161A gene. Exon 5 of the FAM161A gene is located from position 64,974,016 to position 64,974,115 on chromosome 10 (CanFam 2.0).

From the data presented herein, it appears that the location of the SINE insertion near the splice acceptor site of exon 5 results in skipping of exon 5 during mRNA processing and exon splicing. The results suggest that the difference between FAM161A in PRA affected dogs compared to unaffected dogs is the absence of exon 5. This has been shown in FAM161A mRNA transcripts from dogs affected with PRA which predominantly comprise the aberrant FAM161A isoforms lacking exon 5 {FAM161A_f -₅ and FAM161A_S^₅).

In one embodiment, the SINE insertion is less than 500 bp long, such as less than 200 bp long, for example less than 150 bp long, in particular 132 bp long.

In one embodiment, the insertion further comprises a dinucleotide repeat, for example a cytosine-thymine nucleotide repeat. In a further embodiment, the dinucleotide repeat is 8 repeats {i.e. 16 nucleotides) long. In one embodiment, the insertion further comprises a poly(A) tract which is at least 35 bp in length, for example 44 bp or 49 bp in length, in particular 44 bp in length. References herein to a 'polyCA) tract' or 'polyadenine tract' refer to a length of DNA (or RNA) which includes multiple consecutive adenine bases. It can be used to encode a region on mRNA that promotes the assembly of the spliceosome, i.e. the protein complex which carries out RNA splicing during the process of post- transcriptional modification.

In one embodiment, the insertion is about 230 bp long. It will be understood that the term 'about' indicates the value is approximate and may be varied by a reasonable degree. For example, the insertion may be 200 to 250 bp long, for example 220 to 240 bp long, such as 224 to 238 bp long, in particular 238 bp long.

Screening

Progressive retinal atrophy in dogs is an autosomal recessive condition. Thus the progressive retinal atrophy status may be selected from : clear of progressive retinal atrophy, affected by {i.e. having or likely to develop) progressive retinal atrophy, or a carrier of progressive retinal atrophy.

The individual animal tested may or may not be entirely symptomless and\or may be considered to be at risk from progressive retinal atrophy (based on pedigree etc.).

In one embodiment, the method additionally comprises the step of establishing whether or not the canine mammal is heterozygous or homozygous for the genetic variation within the FAM161A gene.

In one embodiment, if the canine mammal is homozygous for the genetic variation within the FAM161A gene, it is prognosed as a canine mammal that will suffer from progressive retinal atrophy. In an alternative embodiment, if the canine mammal is heterozygous for the genetic variation within the FAM161A gene, it is selected as being suitable for breeding with a canine mammal of the same breed which is homozygous for the wild-type FAM161A gene.

In a further alternative embodiment, if the canine mammal is homozygous for the wild-type FAM161A gene, it is selected as being suitable for breeding with a canine mammal of the same breed which is homozygous or heterozygous for the wild-type FAM161A gene.

In one aspect of the invention, the method may include the step of screening a canine mammal for progressive retinal atrophy as described herein, and if the animal is identified as a carrier, selecting it for breeding with an animal which is not a carrier of progressive retinal atrophy i.e. is clear of progressive retinal atrophy and homozygous for the non-mutant, wild-type allele). The ability to identify carriers for breeding purposes is of great importance because progressive retinal atrophy is an extremely debilitating disease which invariably leads to total blindness. The method of the invention may optionally comprise, in addition to detecting genetic variation within the FAM161A gene, the assessment from the same sample for other markers which are linked or associated with other canine disorders. Thus, in one embodiment, the sample is assessed for one or more other markers which are linked or associated with canine disorders.

Therapy

According to a further aspect of the invention, there is provided a method of treating progressive retinal atrophy in a canine mammal, which comprises assessing the progressive retinal atrophy status of a canine mammal by use of a method as defined herein and if the canine mammal is identified as affected by progressive retinal atrophy, treating said canine mammal to prevent or reduce the onset of progressive retinal atrophy. In one embodiment, the method of treating progressive retinal atrophy comprises the following steps:

(a) diagnosing progressive retinal atrophy in a canine mammal according to the method described herein; followed by

(b) administering a medicament to said canine mammal in the event of a positive diagnosis for progressive retinal atrophy.

Gene replacement therapy

As noted above the present inventors have identified a mutation in intron 4 of the FAM161A gene in the DNA which changes the reading frame of the mRNA, in turn producing an isoform that lacks exon 5.

Thus, according to a further aspect of the invention, there is provided a method of treating progressive retinal atrophy in a canine mammal, which comprises increasing the level of non-mutant, wild-type FAM161A gene expression and/or FAM161A gene product activity in the canine mammal .

Normal {i.e. non-mutant) FAM161A gene nucleic acid sequences described above can, for example, be utilized for the treatment of progressive retinal atrophy. Such treatment can be administered, for example, in the form of gene replacement therapy. Specifically, one or more copies of a normal FAM161A gene or a portion of the FAM161A gene that directs the production of a FAM161A gene product exhibiting normal FAM161A gene function, may be inserted into the appropriate cells within a canine mammal in need of the same, using vectors that include, but are not limited to adenovirus, adeno-associated virus, and retrovirus vectors, in addition to other particles that introduce DNA into cells, such as liposomes.

Also included are methods using liposomes either in vivo, ex vivo or in vitro wherein FAM161A gene DNA is delivered to the cytoplasm and nucleus of target cells.

In another embodiment, techniques for delivery involve direct administration of such FAM161A gene sequences to the site of the cells in which the FAM161A gene sequences are to be expressed. Additional methods that may be utilized to increase the overall level of FAM161A gene expression and/or FAM161A gene product activity include the introduction of appropriate FAM161A gene expressing cells, preferably autologous cells, into the canine mammal at positions and in numbers that are sufficient to ameliorate the symptoms of progressive retinal atrophy. Such cells may be either recombinant or non- recombinant. The expression of the FAM161A gene sequences is controlled by the appropriate gene regulatory sequences to allow such expression in the necessary cell types. Such gene regulatory sequences are well known to the skilled artisan. Such cell-based gene therapy techniques are well known to those skilled in the art, see e.g. Anderson, U.S. Pat. No. 5,399,349.

When the cells to be administered are non-autologous cells, they can be administered using well known techniques that prevent a host immune response against the introduced cells from developing. For example, the cells may be introduced in an encapsulated form which, while allowing for an exchange of components with the immediate extracellular environment, does not allow the introduced cells to be recognized by the host immune system. Thus, for example, the invention provides a method of gene therapy wherein one or more copies of a nucleic acid sequence as described herein {e.g. non-mutant FAM161A gene or an active variant thereof) may be inserted into the appropriate cells within the canine mammal, using vectors that include, but are not limited to adenovirus, adeno-associated virus, and retrovirus vectors, in addition to other particles that introduce DNA into cells, such as liposomes.

Example gene therapy vectors for use in the method of this invention include retroviral or episomal vectors expressing particular desired genes under the control of the promoter and/or the supplemental control sequences disclosed herein (see e.g. Axel et a/., U.S. Pat. No. 4,399,216, and Pastan et a/., U.S. Pat. No. 5,166,059, both incorporated herein by reference). Delivery systems as contemplated herein include both viral and liposomal delivery systems (see e.g. Davis et a/., U.S. Pat. No. 4,920,209, incorporated herein by reference). Such gene therapy vectors may incorporate targeting signals to the appropriate membrane or organ. Alternatively, or additionally cell or organelle specific promoters may be used.

The invention also provides such vectors and DNA molecules for use in a method of treatment of progressive retinal atrophy in a canine mammal.

The invention further provides use of such DNA molecules in the preparation of a medicament, for example for the treatment of a canine mammal . Assessment of Genetic Variation

Methods for detecting or assessing genetic variations are reviewed by Schafer and Hawkins, (Nature Biotechnology (1998)16, 33-39, and references referred to therein) and include: allele specific oligonucleotide probing, amplification using PCR, denaturing gradient gel electrophoresis, RNase cleavage, chemical cleavage of mismatch, T4 endonuclease VII cleavage, multiphoton detection, cleavase fragment length polymorphism, E. coli mismatch repair enzymes, denaturing high performance liquid chromatography, (MALDI-TOF) mass spectrometry, analysing the melting characteristics for double stranded DNA fragments as described by Akey et a/. (2001) Biotechniques 30; 358-367. These references, inasmuch as they may be used in the performance of the present invention by those skilled in the art, are specifically incorporated herein by reference.

The assessment of the genetic variation may be carried out on a DNA microchip, if appropriate. One example of such a microchip-system may involve the synthesis of microarrays of oligonucleotides on a glass support. Fluorescently - labelled PCR products may then be hybridised to the oligonucleotide array and sequence specific hybridisation may be detected by scanning confocal microscopy and analysed automatically (see Marshall & Hodgson (1998) Nature Biotechnology 16 : 27-31 , for a review).

Use of nucleic acid probes

The method of detecting or assessing the genetic variation may comprise determining the binding of an oligonucleotide probe to the nucleic acid sample. Thus, in one embodiment, the detection step of the method defined herein is performed by determining the binding of oligonucleotide probes to the nucleic acid sample, wherein the probes comprise all or part of the wild-type or mutant FAM161A gene. In one embodiment, the oligonucleotide probe may be a primer.

In one embodiment, the oligonucleotide probes bind within intron 4 of the FAM161A gene. The oligonucleotide probes may include a mutant probe which specifically binds to mutant DNA, and a wild-type probe which specifically binds to wild-type DNA only.

In a further embodiment, the oligonucleotide probes are:

Forward mutant: 5 '- G G ATC C CTTTATTTG ATTTTA G A A AG - 3 ' (SEQ ID NO : 1);

Forward wild-type: 5'-TCCCTTCCTTTTATTTGATTTTAGAAAG-3' (SEQ ID NO : 2); and

Reverse: 5'-6FAM-CAACAAACACAACCTGAGCAA-3' (SEQ ID NO : 3).

As detailed by the data provided herein, the FAM161A gDNA isoforms are of the following lengths:

Normal, non-mutant, wild-type FAM161A = 135bp

Mutant FAM161A = 137bp

The probe may comprise a nucleic acid sequence which binds specifically to a particular allele of a polymorphism and does not bind specifically to other alleles of the polymorphism . Where the nucleic acid is double-stranded DNA, hybridisation will generally be preceded by denaturation to produce single- stranded DNA. A screening procedure, chosen from the many available to those skilled in the art, is used to identify successful hybridisation events and isolated hybridised nucleic acid. Probing may employ the standard Southern blotting technique. For instance DNA may be extracted from cells and digested with different restriction enzymes. Restriction fragments may then be separated by electrophoresis on an agarose gel, before denaturation and transfer to a nitrocellulose filter. Labeled probe may be hybridised to the DNA fragments on the filter and binding determined. Binding of a probe to target nucleic acid (e.g. DNA) may be measured using any of a variety of techniques at the disposal of those skilled in the art. For instance, probes may be radioactively, fluorescently or enzymatically labeled.

Polymorphisms may be detected by contacting the sample with one or more labeled nucleic acid reagents including recombinant DNA molecules, cloned genes or degenerate variants thereof under conditions favorable for the specific annealing of these reagents to their complementary sequences within the relevant gene.

As is understood by those skilled in the art, a 'complement' or 'complementary' or 'reverse complement' sequence (the terms are equivalent) is one which is the same length as a reference sequence, but is 100% complementary thereto whereby each nucleotide is base paired to its counterpart running in anti-parallel fashion i.e. G to C, and A to T or U.

In one embodiment, the lengths of these nucleic acid reagents are at least 15 to 30 nucleotides.

After incubation, all non-annealed nucleic acids are removed from the nucleic acid :gene hybrid. The presence of nucleic acids that have hybridized, if any such molecules exist, is then detected. Using such a detection scheme, the nucleic acid from the cell type or tissue of interest can be immobilized, for example, to a solid support such as a membrane, or a plastic surface such as that on a microtitre plate or polystyrene beads. In this case, after incubation, non- annealed, labeled nucleic acid reagents are easily removed. Detection of the remaining, annealed, labeled nucleic acid reagents is accomplished using standard techniques well-known to those in the art. The gene sequences to which the nucleic acid reagents have annealed can be compared to the annealing pattern expected from a normal gene sequence in order to determine whether a gene mutation is present. Approaches which rely on hybridisation between a probe and test nucleic acid and subsequent detection of a mismatch may be employed. Under appropriate conditions (temperature, pH etc.), an oligonucleotide probe will hybridise with a sequence which is not entirely complementary. The degree of base-pairing between the two molecules will be sufficient for them to anneal despite a mismatch. Various approaches are well known in the art for detecting the presence of a mis-match between two annealing nucleic acid molecules. For instance, RNase A cleaves at the site of a mis-match. Cleavage can be detected by electrophoresing test nucleic acid to which the relevant probe or probe has annealed and looking for smaller molecules {i.e. molecules with higher electrophoretic mobility) than the full length probe/test hybrid. Other approaches rely on the use of enzymes such as resolvases or endonucleases.

Thus, an oligonucleotide probe that has the sequence of a region of the normal gene (either sense or anti-sense strand) in which mutations and/or polymorphisms associated with the trait of interest are known to occur may be annealed to test nucleic acid and the presence or absence of a mis-match determined. Detection of the presence of a mis-match may indicate the presence in the test nucleic acid of a mutation associated with the trait. On the other hand, an oligonucleotide probe that has the sequence of a region of the gene including a mutation associated with disease resistance may be annealed to the test nucleic acid and the presence or absence of a mis-match determined. The presence of a mismatch may indicate that the nucleic acid in the test sample has the normal sequence, or a different mutant or allele sequence. In either case, a battery of probes to different regions of the gene may be employed.

As discussed above, suitable probes may comprise all or part of the FAM161A gene sequence (or reverse complement thereof), or all or part of a mutant form of the sequence (or reverse complement thereof). The mutant form may contain one or more of the genetic variations described herein.

Those skilled in the art are well able to employ suitable conditions of the desired stringency for selective hybridisation, taking into account factors such as oligonucleotide length and base composition, temperature and so on. Suitable selective hybridisation conditions for oligonucleotides of 17 to 30 bases include hybridization overnight at 42°C in 6X SSC and washing in 6X SSC at a series of increasing temperatures from 42°C to 65°C. One common formula for calculating the stringency conditions required to achieve hybridization between nucleic acid molecules of a specified sequence homology is (Sambrook et al., 1989) : Tm = 81.5°C + 16.6Log [Na + ] + 0.41 (% G+C) - 0.63 (% formamide) - 600/#bp in duplex. Other suitable conditions and protocols are described in Molecular Cloning : a Laboratory Manual : 2nd edition, Sambrook et al., 1989, Cold Spring Harbor Laboratory Press and Current Protocols in Molecular Biology, Ausubel et al. eds., John Wiley & Sons, 1992. Amplification-based methods

The hybridisation of such a probe may be part of a PCR or other amplification procedure. Accordingly, in one embodiment, the detection step is performed by amplifying all or part of the FAM161A gene. The assessment of the genetic variation in the amplification product may then be carried out by any suitable method, e.g. as described herein. An example of such a method is a combination of PCR and low stringency hybridisation with a suitable probe. Unless stated otherwise, the methods of assessing the genetic variation described herein may be performed on a genomic DNA sample, or on an amplification product thereof.

Where the method involves PCR, or other amplification procedure, any suitable FAM161A gene PCR primers flanking the mutation of interest, i.e. the insertion mutation, may be used.

In one embodiment, the amplified region is less than 500 nucleotides in length, such as less than 450 nucleotides in length, in particular less than 400, especially 200 to 400 nucleotides in length. In one embodiment, the detection step is performed by amplifying all or part of exon 5, i.e. nucleotides 64,974,016 to 64,974,115 on chromosome 10, of the FAM161A mRNA. In one embodiment, the detection step is performed by use of primers which flank and/or include nucleotides 64,974,016 to 64,974,115 on chromosome 10 mRNA (CanFam 2.0).

In one embodiment, the detection step is performed by amplifying all or part of intron 4, i.e. nucleotides 64,974,116 to 64,975,764 on chromosome 10, of the FAM161A gene.

In one embodiment, the detection step is performed by use of primers which flank and/or include nucleotides 64,974,116 to 64,975,764 on chromosome 10 (CanFam 2.0).

An oligonucleotide for use in nucleic acid amplification may be about 30 or fewer nucleotides. Generally specific primers are upwards of 14 nucleotides in length, but are suitably 15-25 inclusive, more preferably 18-20. Those skilled in the art are well versed in the design of primers for use processes such as PCR. Various techniques for synthesizing oligonucleotide primers are well known in the art, including phosphotriester and phosphodiester synthesis methods.

In one embodiment, assessment of the FAM161A gene will establish whether or not the individual animal is heterozygous or homozygous for the specific length variant in this region.

Nucleic acids for use in the methods of the present invention, such as an oligonucleotide probe and/or pair or selection of amplification primers, may be provided in isolated form and may be part of a kit, e.g. in a suitable container such as a vial in which the contents are protected from the external environment. The kit may include instructions for use of the nucleic acid, e.g. in PCR and/or a method for determining the presence of nucleic acid of interest in a test sample. A kit wherein the nucleic acid is intended for use in PCR may include one or more other reagents required for the reaction, such as polymerase, nucleotides, buffer solution etc. The nucleic acid may be labelled. A kit for use in determining the presence or absence of nucleic acid of interest may include one or more articles and/or reagents for performance of the method, such as means for providing the test sample itself, e.g. a swab for removing cells from the buccal cavity or a syringe for removing a blood sample (such components generally being sterile).

The various embodiments of the invention described above may also apply to the following : a means for prognosing progressive retinal atrophy in a canine mammal; a prognostic kit comprising such a means; and the use, in the manufacture of means for prognosing progressive retinal atrophy in a canine mammal of sequences {e.g., PCR primers) to amplify a region of the FAM161A gene as described herein.

The invention further provides oligonucleotides for use in probing or amplification reactions, which may be fragments of the FAM161A gene.

According to a further aspect of the invention, there is provided oligonucleotide probes for use in a method of prognosing progressive retinal atrophy in a canine mammal, wherein said oligonucleotide probes detect a genetic variation within the FAM161A gene and are as defined herein.

Suitable polymerase chain reaction (PCR) methods are reviewed, for instance, in "PCR protocols; A Guide to Methods and Applications", Eds. Innis et al., 1990, Academic Press, New York, Mullis et al., Cold Spring Harbor Symp. Quant. Biol ., 51 : 263, (1987), Ehrlich (ed), PCR technology, Stockton Press, NY, 1989, and Ehrlich et al., (1991) Science 252 : 1643-1650). PCR comprises steps of denaturation of template nucleic acid (if double-stranded), annealing of primer to target, and polymerisation.

An amplification method may be a method other than PCR. Such methods include strand displacement activation, the QB replicase system, the repair chain reaction, the ligase chain reaction, rolling circle amplification and ligation activated transcription. For convenience, and because it is generally preferred, the term PCR is used herein in contexts where other nucleic acid amplification techniques may be applied by those skilled in the art. Unless the context requires otherwise, reference to PCR should be taken to cover use of any suitable nucleic amplification reaction available in the art.

In one embodiment, "Amplified Fragment Length Polymorphism" (AFLP) may be carried out using primers devised on the basis of the sequences disclosed herein. Analysis of the products can be carried out using e.g. by gel electrophoresis, capillary electrophoresis.

In one embodiment, described in the Examples hereinafter, the region of DNA that contains the mutation is amplified using PCR and the length of the resulting fragment of DNA is measured.

Examples of results from the genotyping assay are shown below. Sequencing

The genetic variation may be assessed or confirmed by nucleotide sequencing of a nucleic acid sample to determine the presence of the genetic variation. The identity may be determined by comparison of the nucleotide sequence obtained with the native, non-mutant, wild-type sequence.

Nucleotide sequence analysis may be performed on a genomic DNA sample, or amplified part thereof, or RNA sample as appropriate, using methods which are standard in the art.

Where an amplified part of the genomic DNA sample is used, the genomic DNA sample may be subjected to a PCR amplification reaction using a pair of suitable primers. In this way the region containing a particular polymorphism or polymorphisms may be selectively amplified (PCR methods and primers are discussed in more detail herein). The nucleotide sequence of the amplification product may then be determined by standard techniques. Other techniques which may be used are single base extension techniques and pyrosequencing.

Mobility based methods

The assessment of the genetic variation may be performed by single strand conformation polymorphism analysis (SSCP). In this technique, PCR products from the region to be tested are heat denatured and rapidly cooled to avoid the reassociation of complementary strands. The single strands then form sequence dependent conformations that influence gel mobility. The different mobilities can then be analysed by gel electrophoresis.

Assessment may be by heteroduplex analysis. In this analysis, the DNA sequence to be tested is amplified, denatured and renatured to itself or to known wild-type DNA. Heteroduplexes between different alleles contain DNA "bubbles" at mismatched basepairs that can affect mobility through a gel. Therefore, the mobility on a gel indicates the presence of sequence alterations.

Restriction site based methods

Where a mutation creates or abolishes a restriction site, the assessment may be made using RFLP analysis. In this analysis, the DNA is mixed with the relevant restriction enzyme (i.e., the enzyme whose restriction site is created or abolished). The resultant DNA is resolved by gel electrophoresis to distinguish between DNA samples having the restriction site, which will be cut at that site, and DNA without that restriction site, which will not be cut.

Where the mutation does not create or abolish a restriction site the mutation may be assessed in the following way. A mutant PCR primer may be designed which introduces a mutation into the amplification product, such that a restriction site is created when one of the polymorphic variants is present but not when another polymorphic variant is present. After PCR amplification using this primer (and another suitable primer), the amplification product is admixed with the relevant restriction enzyme and the resultant DNA analysed by gel electrophoresis to test for digestion. Kits

According to a further aspect of the invention, there is provided a kit for use in a method of prognosing progressive retinal atrophy in a canine mammal, wherein said kit comprises:

(a) oligonucleotide probes which detect a genetic variation within the

FAM161A gene and are as defined herein; and

(b) means for providing a test sample from the canine mammal.

A kit for use in determining the presence or absence of nucleic acid of interest may include one or more articles and/or reagents for performance of the method, such as means for providing the test sample itself, e.g. a cheek swab (such components generally being sterile). Such a kit may also include instructions for use. The following studies illustrate the invention :

MATERIALS AND METHODS

Sample processing

The diagnosis of individual dogs was determined by veterinary ophthalmologists independently, or through the BVA/KC/ISDS (British Veterinary Association/Kennel Club/International Sheep Dog Society) Eye Scheme in the UK. Cases were defined as dogs diagnosed as affected with Progressive Retinal Atrophy (PRA) i.e. displaying ophthalmascopic signs of PRA including tapetal hyperreflectivity and vascular attenuation. Controls were those free of inherited eye disease of any kind, and at least 4 years old at the time of examination for the genome wide association analysis or any age for subsequent investigations.

Blood samples were collected into EDTA tubes and genomic DNA was extracted from whole blood using a Nucleon Genomic DNA Extraction Kit (Tepnel Life Sciences, Manchester, UK), according to the manufacturer's instructions. For samples collected as buccal mouth swabs, DNA was extracted using a QIAamp^® DNA Blood Midi Kit (Qiagen, West Sussex, UK). A canine retinal tissue sample from a dog of unknown breed and free of PRA was taken post mortem, with the owner's consent. RNA was extracted using an RNeasy Protect Mini Kit (Qiagen, West Sussex, UK) according to the manufacturer's instructions.

Blood samples from two dogs (a Tibetan Spaniel (TS) with PRA and homozygous for the SINE insertion, and from a dog of unknown breed but free of PRA) were collected into EDTA tubes. RNA was extracted using the PerfectPure RNA Blood Kit (5 Prime, USA) according to the manufacturer's instructions.

Genome-wide association mapping

Canine SN P20 BeadChips (Illumina) were used to obtain genotype calls for 22,362 single nucleotide polymorphisms (SNPs) using DNA from 22 TS PRA cases and 10 TS controls and GWA analysis was conducted using the software package PLINK [37] . After removing SN Ps with a minor allele frequency less than 5% and missing genotype calls greater than 10% from the analysis, a final data set of 15,674 markers remained. Sample call rate was greater than 99.7% for all samples. Identity-by-state (IBS) clustering and Cochran-Mantel-Haenszel (CMH) meta-analysis with PLINK were used to examine and adjust for population stratification. A mixed model analysis using Fast Mixed Model (FMM) was also undertaken to correct for population stratification. As a correction for multiple testing, the GWA analyses using the Max(T) permutation procedure in PLIN K was repeated (100,000 permutations). P-values generated before multiple testing correction are denoted by p_raw, while those generated after are denoted by

Pgenome- Microsatellite marker genotyping

Microsatellite markers within the associated region were genotyped in 18 cases and 5 controls used in the GWA investigation (Table 1). Microsatellite markers were identified by searching the reference sequence (CanFam2.0) for dinucleotide repeats with a total length of at least twenty nucleotides (sequences available upon request). Primers flanking each marker were designed using Primer3 so that the resulting products would be between 200 and 400 base pairs in size [38] . A tail of 18 bp (5'- TGACCGGCAGCAAAATTG-3' (SEQ ID NO : 4)) was added to the 5' end of the left primer of each pair. After amplification with the third, fluorescent primer complementary to the tail (5'-FAM- CAATTTTGCTGCCGGTCA-3' (SEQ ID NO : 5)), the products were separated by size on a 3100 or 3130x1 Genetic Analyzer (Applied Biosystems) and the data analysed and alleles assigned to each sample with the GeneMapper software package (Applied Biosystems). Visual inspection of SNP and microsatellite marker genotypes and haplotypes across the region was performed to define a homozygous critical region.

AHT-C-P1603PCT

26

TABLE 1: Primers for microsatellite genotyping

AHT-C-P1603PCT

27

AHT-C-P1603PCT

28

AHT-C-P1603PCT

29

(TAIL = TGACCGGCAGCAAAATTG (SEQ ID NO : 4))

Next generation sequencing

Genomic DNA (3 μς) from 10 TS dogs (four PRA-affected, two obligate carrier and four PRA-clear) was used to prepare libraries for sequencing, using the SureSelectXT Custom MP4 Kit (Agilent Technologies). Each kit contained a custom capture library of 40,473 biotinylated RNA baits, 120 bp in length and designed based on the CanFam2 reference sequence using Agilent Technologies eArray tool [39] . Baits were designed to give 2x coverage and exclude repeat- masked regions, resulting in coverage of 54.5% (2.72/5 Mb) of targeted regions. Target enrichment was performed according to the manufacturer's instructions. Initial shearing of genomic DNA using a Covaris S220 and quality assessment of the final library using a 2100 Bioanalyser was undertaken by The Eastern Sequence and Informatics Hub (EASIH, University of Cambridge). The quantity of the captured library was assessed by quantitative PCR using the KAPA Library Quantification Kit for the Illumina Genome Analyzer Platform (KAPA Biosystems), according to the manufacturer's instructions.

Paired-end sequencing resulting in 51 bp reads was conducted in a single lane on an Illumina HiSeq 2000, by the High Throughput Group (HTG) at the Welcome Trust Centre for Human Genetics, University of Oxford, UK. Sequence reads were aligned with the CanFam2 canine reference sequence using BWA [40], variant (indel and SN P) called were made using GATK [41] and aligned reads were visualised using the Integrative Genomics Viewer (IGV) [42] . Variants considered candidates for further investigation were those that occurred in splice sites or could affect splicing, or resulted in non-synonymous changes to a protein, and were homozygous in PRA cases, heterozygous in obligate carriers and homozygous for the wild-type allele in controls.

Primers

The exon-intron boundaries of canine FAM161A were defined by producing ClustalW [43] alignments using the Ensembl predicted canine transcripts (ENSCAFG00000003079) and available known mouse (NSMUSG00000049811) and human (ENSG00000170264) Ensembl transcripts. Primer3 [38] was used to design all primers (Table 2), fluorescent and non-fluorescent (IDT, Glasgow, UK). These included primers in the exons for the amplification and sequencing of cDNA; in the introns flanking exons one and five for the amplification and sequencing of these exons of FAM161A in genomic DNA; fluorescent primers flanking the SINE insertion to determine the length of the insertion; and allele- specific primers to detect the presence or absence of the insertion. Amplification products generated using fluorescent primers were used for subsequent fragment length polymorphism detection using an ABI 3130x1 DNA Analyzer and GeneMapper^® Software (Applied Biosystems, Inc., [ABI], Foster City, CA).

TABLE 2: Primers for sequencing and genotyping

Sanger Sequencing

FAM161A complimentary DNA (cDNA) was generated using SuperScript^®II Reverse Transcriptase (Invitrogen) according to the manufacturer's instructions. The region containing the SINE insertion was amplified from genomic DNA (gDNA) and the entire gene from cDNA, by polymerase chain reaction (PCR) using HotStarTaq Plus DNA Polymerase (Qiagen). PCR products were purified using Multiscreen HTS-PCR filter plates (Millipore). Amplification products were sequenced on an ABI 3130x1 DNA Analyzer using BigDye Terminator v3.1 (Applied Biosystems) and sequence traces were assembled, analyzed and compared using the Staden Package [44].

AFLP and variant genotyping

To further investigate the variant (FAM161A_c. i₇5₈-i5 _i758-i6in_S238), the 32 TSs (22 cases and 10 controls) that participated in the GWA study were genotyped using the allele-specific fluorescent primers described above. The variant was analysed for association with PRA and compared with the most associated SNP markers, BICF2P582923 and BICF2G630416812, using the software package PLIN K [37] . The suggestive causative mutation for PRA3 in intron 4, FAM 161A_c.i₇₅₈-i5 i₇ss- i6ins238, was then screened in 247 TSs. The panel of 247 TSs (including the 80 DNA samples already sequenced), was made up of 35 PRA cases, 16 obligate carriers, 100 clear dogs and 96 dogs with unknown clinical status. In addition, samples from 99 dogs representing two breeds (23 Lhasa Apsos and 76 Tibetan Terriers) that are closely related to the TS breed were also included in the mutation screening.

RESULTS

Genome-wide Association Mapping

Genome-wide association (GWA) analysis of genotyping data from 32 Tibetan Spaniel (TS) dogs (22 cases and 10 controls over the age of four when last examined) genotyped with 15,674 SN Ps revealed a genome-wide significant association on chromosome 10 (CFA10; p_raw = 1.77xl0^"7, p_genome = 0.004). Two SNP markers 1.86 Mb apart (BICF2P729624 at 62.0 Mb and BICF2S23250878 at 63.86 Mb) were equally the most associated with PRA. Identity-by-state (IBS) clustering using genome-wide SNP marker data confirmed the presence of population stratification with a genomic inflation factor > 1 (λ = 1.69 based on the median χ2). The inflation factor was reduced to an acceptable level (λ = 1.06) after correcting for population stratification with a Cochran-Mantel- Haenszel (CMH) meta-analysis. While the signal on CFAIO (p_raw = 2.01xl0^"5, Figure 1A) dropped below the level of Bonferroni-corrected significance, the permutation-corrected signal remained statistically associated (p_genome = 0.014). Alternative analysis of the data using Fast Mixed Model (FMM) to correct for population stratification revealed similar results, with the strongest signal on CFAIO (p_raw = 5.67xl0^"5, Figure IB) and a reduced inflation factor (λ = 1.27). Using both CMH and FMM corrections the most associated SNP was BICF2S23422025 at 66.74 Mb. The signal on CFAIO was at least 10-fold stronger than a number of signals seen on other chromosomes including 1, 8 and 38. While only a single SNP (BICF2S23422025 at 66.74 Mb: p_genome = 0.014) was statistically significantly associated, the signal on CFAIO extended from approximately 62 to 67 Mb (Figure 1C).

Haplotype and homozygosity analysis

A haplotype homozygous in cases, but not in controls, could not be easily identified through homozygosity analysis. The most highly-associated SNP in CMH and FMM analyses, BICF2S23422025 (p_genome = 0.014) is homozygous (A/A) in most of the cases (19/22), but also in 3/10 controls. The SNP20 BeadChip contains a relatively small number of SN Ps resulting in a low genotyping resolution - one SNP approximately every 114 kb, if all SN Ps are informative. In the case of the TS cohort, only 15,674 SNPs were informative, resulting in 1 SN P approximately every 159 kb on average. This made it difficult to identify homozygous haplotypes from the SNP data alone and additional microsatellite markers from the region were therefore genotyped to provide additional haplotype information (Figure 2). Due to sample availability constraints (low DNA quantity) only 18 of the original 22 cases and five of the original ten controls (one of which is an obligate carrier) used in the GWA study were used in the microsatellite marker genotyping study. Even with the increased resolution it was not possible to define a haplotype that was homozygous in all cases. However, a broad critical region of 3.794 Mb, from 63.935 Mb to 67.729 Mb on CFAIO was identified that is almost completely homozygous in most of the cases (12/18) and none of the controls, and heterozygous in the obligate carrier. This region contains 31 genes, 29 of which have human orthologues. At the time this work was undertaken, none of the genes in the region could be identified as strong functional candidates. Due to the lack of identifiable candidate genes, sequencing of the whole critical region with next generation sequencing was chosen for further investigation of the TS PRA locus.

Sequencing

To identify potential disease-causing mutations, the inventors undertook targeted re-sequencing of the critical region using 10 samples (four affected, two obligate carrier and four normal dogs). Repetitive DNA elements, making up approximately 46% of the regions, were masked during the design of custom RNA baits and as a result approximately 56% of the 5 Mb targeted region was enriched and sequenced. More than 193 million reads were generated across all 10 samples (representing a 9.9 Gb dataset), of which 72% were mapped to the targeted regions on CFA10. The average read depth across the targeted region for each sample ranged from 102x to 174x, and approximately 65% of the region covered by baits was sequenced with at least 30x coverage. More than 19,000 SNPs and 3,700 indels were identified when compared with the CanFam2 reference sequence. Of these 194 SNPs and 81 indels segregated with the phenotype, but none of these variants were predicted to alter the protein product. Visual analysis of sequence data in IGV revealed 16 additional variants (larger insertion and deletions not identified using the NGS analysis pipeline), of which seven segregated with the phenotype, but only one was in or near an exon. This variant was an insertion flanked by a 14 bp repeat motif (Figure 3), visualised in IGV as an increase in the read depth. The length of the inserted sequence is longer than the length of the NGS reads (>50 bp) and the precise sequence of the insertion could therefore be only partly determined (Figure 3B). Only this variant, which was predicted to be located near a splice acceptor site of the FAM161A gene (CFA8 : 64,974,130), could potentially alter the protein product, by interfering with exon splicing.

The full sequence of the insertion was determined by Sanger sequencing using primers flanking exon 5 of FAM161A, including the insertion site in gDNA from 80 TS dogs (29 affected with PRA, 10 obligate carriers and 41 unaffected). Using agarose gel electrophoresis, a single band of the expected size (720 bp) was visible for 40 unaffected samples, while a band approximately 230 bp larger (approximately 950 bp) was visible for 17 of the PRA affected samples and none of the carrier or unaffected samples. Carriers of the insertion were more difficult to identify from gel electrophoresis alone (the larger band was not as bright as the smaller band, and sometimes not visible at all), presumably due to preferential amplification of the shorter DNA fragment lacking the insertion. However, carriers were identifiable by Sanger sequencing. Sequencing of the approximately 230 bp insertion revealed that it contains a 132 bp SIN E, a retroposon that is distributed widely throughout the canine genome [20] . As is characteristic, the SINE is followed at the 3' end by a dinucleotide repeat, (CT)8, and a poly(A) tract (interrupted by the occasional T) at least 45 bp in length. The nucleotides at the 3'-end of the poly(A) tract are duplicated at the 5'-end of the SINE (Figure 4). The number of adenine nucleotides that comprise a portion of the poly(A) tract (underlined in Figure 4A) could not be determined accurately due to difficulties amplifying homopolymers with synthetic taq polymerases, specifically polymerase slippage along the poly(A) tract. However, based on the sequence traces, there appear to be 35-50 adenine nucleotides. Poly(A) Tract Length

An amplified fragment length polymorphism (AFLP) assay using fluorescent primers flanking the SIN E insertion was used to further investigate the length of the insertion and therefore the poly(A) tract. As a result of Taq polymerase slippage on the poly(A) tract, amplification of the SINE insertion resulted in multiple products (viewed as fluorescent peaks) varying in size, creating a "hedgehog" effect. The SINE insertion was amplified from 21 dogs that were either homozygous (n = 14) or heterozygous (n = 7) for the SINE insertion, as determined by sequencing. For most dogs (15/21), the highest peak was at 391 bp. The size of the wild-type amplicon is 153 bp, and this suggests the SINE insertion is 238 bp in size. Therefore, the poly(A) tract (represented by the underlined text in Figure 4A) is 44 nucleotides in length. In 5/21 samples assayed the highest peak was at 396 bp, suggesting a poly(A) tract five nucleotides longer i.e. 49 nucleotides. One dog carried a single copy of each of the 391 and 396 bp alleles. The age at diagnosis was known for 12 of the dogs assayed that were clinically affected with PRA. The age of onset of PRA in dogs with only the 391 bp allele ranged from 2.9 to 10.2, but there was insufficient data for the 396 bp allele. These data are consistent with the length of the poly(A) tract having no effect on the age at diagnosis.

Transcript evaluation

In humans FAM161A occurs in two main isoforms, full-length (FAM161Au) and short {FAM161Ash), formed by alternative splicing of exon 4 (Figure 5) [18,19] . Four incomplete canine FAM161A isoforms were predicted by Ensembl, but alignment with human and mouse genomic and coding sequences revealed several inconsistencies and possible errors in the prediction of canine intron- exon boundaries (Figure 5A-C). Most of the coding sequence of the FAM161A retinal transcripts, from both the main isoforms (FAM161Au and FAM161A_Sh), was successfully sequenced in a healthy dog, excluding the first 46 nucleotides of the coding sequence (Figure 5D and E). Sequencing revealed that both isoforms are transcribed in the canine retina. In addition, intron-exon boundaries are identical to those of the human and mouse, which is in conflict with the boundaries predicted by Ensembl genebuild for the canine gene. Canine retinal mRNA sequence differed from CanFam2.0 reference sequence in exon 1, and two variants were identified : the first was a SN P that changes an alanine residue to an aspartic acid residue; the second was a 6 bp insertion resulting in the in- frame insertion of two alanine residues (Figure 5E). To assess the prevalence of these variants, genomic DNA from 43 TSs and 76 dogs comprising 31 breeds were sequenced. The SNP variant was present in the homozygous and heterozygous state in both cohorts i.e. TSs and multiple other breeds). Eight dogs from five breeds, excluding TSs, were heterozygous for the insertion variant and none were homozygous. It is unclear what effect, if any, these variants have on the protein. Sequencing of the full 5' and 3' UTRs and the beginning of exon 1 was unsuccessful, probably due to high GC content. Sequencing revealed that canine FAM161Au contains 716 amino acids (Genbank accession no KF177335) and FAM161A_S contains 660 amino acids (Genbank accession no KF177336), with predicted molecular weights of 83kDA and 76kDa respectively. The SINE insertion occurs near the acceptor splice site of intron 4 i.e. near the boundary of intron 4 and exon 5 (FAM161Au : c. l758-15_1758- 16ins238; FAM161A_sh C.1590-15_1590-16ins238) (Figure 5D).

Blood Transcripts

The location of the SINE insertion near the splice acceptor site of exon 5 suggests exon splicing may be affected, possibly resulting in the skipping of exon 5. To assess this hypothesis, mRNA transcripts were compared between a TS dog homozygous for the SINE insertion and two dogs of unknown breed. In the absence of suitable retinal tissue, RNA was purified from the blood of the affected and one of the unaffected dogs, while retinal tissue was available from the other unaffected dog. Primers in exons 3 and 6 were used to amplify across exons 4 and 5. A number of products were produced for all three samples, each of which was individually sequenced (Figure 6) :

Bands 4 and 6 comprised the 421 bp amplicon (FAM161An), and bands 5 and 8 comprised the 253 bp amplicon {FAM161A_S^). Both of these isoforms were detectable in normal blood and retina, but not affected blood.

Bands 2 and 7 comprised the 321 bp amplicon (FAM161Au-₅) and bands 3 and 9 comprised the 153 bp amplicon {FAM161A_s^-₅). Both of these amplicons were detectable in affected blood. Interestingly these bands were also detectable in normal retina, albeit at lower levels than FAM161A_f and FAM161A_S , but not unaffected blood. Band 1 comprised an amplicon containing FAM161Au-₅ and intron 3, which could be a result of gDNA contamination or incomplete or inaccurate exon splicing.

These results suggest that the difference between FAM161A in PRA3 affected and normal blood is the absence and presence respectively of exon 5, supporting the exon-skipping hypothesis.

Mutation screening

All 32 TS dogs (22 cases and 10 controls) that participated in the GWA study were screened for the SINE insertion

using allele- specific fluorescent primers, to confirm the association of this variant with PRA and compare it with two of the most highly associated SNP markers, BICF2S23422025 and BICF2S23250878. FAM161A_C 1758-15_1758-16ins238 Showed significant allelic association with PRA (p_raw = 5.03xl0^"7). The SNP markers also showed significant allelic association, although BICF2S23422025 was less associated (p_raw = 6.28xl0^"7). However, BICF2S23250878 was more highly associated (p_raw = 1.77xl0^"7) than FAM161A_C. 1758-15_1758-16ins238_/ DUt th IS 03Π be attributed to two PRA cases that are heterozygous for the SNP i.e. carry the minor allele) but homozygous for the wild-type FAM161A allele. Fifteen out of 22 PRA cases and none of the controls were homozygous for FAM161A_c._i75?,._i5 1753- i6ins238- Analysis of the segregation of FAM16- _c. i758-i5_i758-i6ins238 with PRA within a family of 49 dogs, including seven cases, indicates that the form of PRA associated with this variant is recessive and fully penetrant. The form of PRA that is associated with FAM16- _c. i758-i5_i758-i6ins238 is known hereafter as PRA3.

To confirm that the variant is not a commonly occurring polymorphism in this breed, we screened 215 additional TS dogs, resulting in a total of 247 TSs tested for FAM161Ac i758-i5_i758-i6ins238 (Table 3). Of the 35 PRA cases used in the study 22 (62.9%) were homozygous for

_i758-i6ins238 (FAM161A^~/~) and all 116 dogs known to be clinically free of PRA at their last eye examination, including 16 obligate carriers of PRA, were either carriers of the mutant allele (14.7%; FAM161A^+/~) or homozygous for the wild type allele (85.3%; FAM161A^+/+). PRA3 therefore accounts for the majority of cases of PRA in our TS cohort.

TABLE 3: PR A3 genotypes and PRA clinical status for 247 TSs

The wild-type allele is represented by "+" and the mutant allele by To determine whether FAM16J^_c. i758-i5_i758-i6ins238 is associated with PRA in related breeds we screened a further 99 dogs from two closely related breeds most likely to share polymorphisms with the TS breed. These were 23 Lhasa Apsos (LA) and 76 Tibetan Terriers (TT), including nine LAs and 12 TTs affected with PRA. All 23 LA dogs, including nine PRA cases, were homozygous for the wild-type allele (FAM161A^+/+). PRA3 is therefore absent from this LA cohort, but as the number of dogs tested was small, it cannot be eliminated entirely as a form of PRA in the breed. Of the 12 TTs with PRA, four were homozygous for FAM16- _c. i758-i5_i758-i6ins238 (FAM161A^{' '}) , while the remaining eight PRA cases were either carriers (FAM161A^+/~; n = l) or homozygous wild-type (n=7). In addition, all TTs known to be free of PRA (n = 10) were homozygous for the wild-type allele (79.7%). PRA3 is therefore present in the TT breed.

DISCUSSION

Using a GWA mapping and homozygosity analysis approach, a novel 3.794 Mb locus on chromosome 10 that is associated with PRA in the TS was identified. The entire critical region was sequenced and a single provocative variant was identified in the FAM161A gene. FAM161A was subsequently identified as a strong candidate causal locus.

Sequencing of FAM161A from healthy retinal mRNA served three purposes: Firstly, it confirmed the presence of FAM161A mRNA transcripts in the normal canine retina. Secondly, it revealed that the intron-exon boundaries predicted by genebuild for FAM161A in the dog are incorrect for three exons. They are instead identical to the human and mouse boundaries. Thirdly, as is the case in humans, canine FAM161A is alternatively spliced to produce two isoforms, one containing and one lacking exon 4 (FAM161Au and FAM161A_Sh, respectively).

Sanger sequencing and preliminary qPCR results indicate that FAM161A mRNA transcripts in retinal tissue and blood from dogs not affected with PRA comprise predominantly the wild-type FAM161A isoforms (FAM161A_f and FAM161A_Sh, Figure 6). Conversely, FAM161A mRNA transcripts in blood from a dog affected with PRA3 i.e. homozygous for

comprises predominantly the aberrant FAM161A isoforms lacking exon 5 (FAM161Au-₅ and FAMieiAsh s)- This supports the hypothesis that the SINE insertion results in skipping of exon 5 during pre-mRNA splicing in blood. While it is likely that FAM161Ar i758-i5_i758-i6ins238 has the same effect of aberrant splicing in other tissues, the possibility that tissue-specific splicing negates this effect in the retina cannot be excluded. Further investigation using retinal tissue from a dog with PRA3 is necessary to substantiate the hypothesis of alternative splicing. Interestingly, aberrant FAM161A isoforms

and FAM161A_s -₅) were also present in retinal tissue from a dog not affected with PRA, albeit at much lower levels than the wild-type isoforms. These are most likely a result of naturally-occurring alternative splicing, which is a common occurrence. At least 74% of human multi-exon genes are alternatively spliced [21] and up to 30% of alternative transcripts contain premature termination codons [22] . These are usually targets of nonsense-mediated decay (NMD), although Lewis et al. observed that 4.3% of RefSeq mRNAs {i.e. experimentally identified mRNAs that have not been degraded) are truncated by more than 50 amino acids [22] . While these aberrant proteins may well be expressed in healthy retinal tissue, it is clear from PCR/electrophoresis data presented here that they are a minor product compared with the normal, functional protein. The pre-mRNA splicing mechanism requires at least three consensus intronic sequences for optimal function. One of these is the 3' consensus sequence 6PyNCAG (where Py is a pyrimidine base, N is any base and Pu is a purine base) of the acceptor site and another, the branch point sequence (BPS), is the site of lariat formation [23] . In eukaryotes the latter is typically, but not always, 20 to 50 nucleotides upstream of the splice junction. The consensus sequence of the BPS, to which the U2 component of the spliceosome binds, is also variable (PyXPyTPuAPy), although the adenine base is of primary importance for lariat formation [24,25] . There is no sequence within 50 nucleotides of the FAM161A intron 4-exon 5 splice site that corresponds to the BPS consensus sequence. However, a putative BPS located 76 nucleotides from the splice site, does correspond to the consensus sequence (Figure 7) . While the SINE insertion does not affect the AG sequence of the acceptor site it will push the BPS beyond its optimal range from the acceptor site, which is likely to be the cause of aberrant splicing of exon 5 in affected dogs. Similar instances of aberrant splicing due to SINE insertions that are associated with canine traits have been reported : An insertion 35 bp upstream of an acceptor site of the HCRTR2 (hypocretin (orexin) receptor 2) gene has been associated with canine narcolepsy in the Doberman breed [26], and an insertion nine bp upstream of an acceptor site of the SILV (a.k.a PMEL; premelanosome protein) gene has been associated with the merle pigmentation pattern in multiple dog breeds [27] . The authors of both studies hypothesised that the intronic SINE insertion near a splice acceptor site displaced a putative lariat BPS, resulting in skipping of the adjacent exon during mRNA splicing. Lin et al. presented data in support of this hypothesis: HCRTR2 cDNA from dogs homozyhous for the SINE insertion associated with canine narcolepsy lacked the sequence corresponding to the entire fourth exon [26] .

The broad range of ages-at-diagnosis observed in PRA-affected dogs homozygous for FAM16- _c. i758-i5_i758-i6ins238 suggests there may be a great deal of variation in the age-of-onset or severity of PRA3. This could be due to a variable poly(A) tract length. Longer poly(A) tracts in human LI long interspersed nuclear elements (LIN Es) have been associated with disease in humans [28] and the longer SINE in the SILV gene results in the merle phenotype in dogs, while the shorter SINE does not [27] . However, Miyadera et al. reported that variable poly(A) tract length in the RPGRIP1 SINE insertion associated with cone-rod degeneration (CORD1) in Miniature Longhaired Dachshunds did not correlate with variable phenotype [29] . While the precise length of the poly(A) tract in the FAM161A SINE insertion is unknown, data presented suggests the length of the poly(A) tract does vary, with two alleles identified (A44 and A49). However, as the predominant allele (A44) segregates in dogs that were diagnosed with PRA at between 2 and 10 years of age, it is unlikely that the poly(A) tract length variation has much effect on the PRA phenotype or severity.

In order to further test the validity of the insertion variant, FAM161A_C.₁₇₅₈- 15 1758- i6ins238, we screened 247 TSs for the variant (Table 3). We found that 62.9% of the PRA cases, 56.3% of the obligate PRA carriers and 100% of clinically unaffected dogs (which could be clear of the variant or carry a single copy) have FAM161A genotypes that are concordant with their clinical status. There are two groups of dogs with genotypes discordant with their phenotypes. The first comprises two dogs that are homozygous for the variant and have not been diagnosed with PRA. Clinical information pertaining to one of these dogs was unavailable, although it is known to have had at least one PRA-affected sibling. The other dog had not been seen by an ophthalmologist but its owner reported no significant loss of sight by the time it died. However, it had lost one eye in an accident and developed a cataract in the other eye around nine years of age, which could have been secondary to PRA. The observation that 91.7% (22/24) of dogs homozygous for FAM161A_c, ijs8-i5 _i758-i6ins238 {i.e. FAM161X'-) have developed PRA suggests the variant is fully penetrant, or nearly so. The inheritance observed in a family of 49 dogs (seven cases) is supportive of a recessive mode. The second group of discordant dogs comprises 13 PRA-affected dogs that are not homozygous for FAM16- _c. i758-i5_i758-i6ins238 and seven obligate carriers do not carry FAM16J^_c. i758-i5_i758-i6ins238- I is formally possible that the variant has a dominant mode of inheritance with incomplete penetrance, or complex trait or compound heterozygote effects. However, as heterogeneity of PRA has been seen in other breeds [30,31] it is more likely that additional loci are responsible for the discordant cases.

Anecdotal evidence of their shared origins in Tibetan monasteries suggests that TTs and LAs are the most closely related breeds to the TS, and as a result these are the breeds most likely to share the PRA3 variant. Screening of 76 TT and 23 LA dogs, including 12 and nine PRA cases respectively, revealed that the variant is present in TTs, but is absent from the LAs screened. FAM161A genotypes of 33.3% of the TT PRA cases and 100% of clinically unaffected dogs (which could be clear of the mutation or carry a single copy) were concordant with their clinical status. Interestingly, two of the eight PRA-affected TTs that were not homozygous for FAM161A_c, ijs8-i5 _i758-i6ins238 {i.e. FAM161A^+/- or FAM161A^+/+), were in fact homozygous for the mutation associated with RCD4 [30] . It is therefore likely that PRA in the remaining six cases in the breed is caused by a third unknown mutation. The allele frequency of FAM16- _c. i758-i5_i758-i6ins238 in both breeds is not unusually high, even for a recessive disease allele. Estimates for both breeds indicate that up to one in 31 TS dogs and one in 48 TT dogs are affected with this form of PRA. However, these frequencies are likely to be inflated, as samples collected for research are unlikely to be representative of the wider population.

FAM161A encodes the family with sequence similarity 161, member A protein. The gene occurs in two main isoforms, that differ by the presence or absence of exon 4

and FAM161A_s , respectively) [18] . Both isoforms are expressed in multiple tissues including the retina and testes, and at lower levels in the heart, liver, kidney, brain, muscle, lung and thyroid gland [19]. Specifically, the protein has been localised to the connecting cilium and basal body in the inner segment of rod and cone photoreceptor cells, and to the basal body and centrosome of ciliated cells of different origins [32,33] . FAM161A has been shown to interact with the CRX (Cone-rod homeobox-containing) transcription factor [19] and Lebercilin [32], both of which have also been implicated in retinal degeneration in humans [34,35] . Only a single evolutionary conserved domain (UPF0564) has been identified, which is vital for binding to and stabilising microtubules [18,33] . This region is also required for homotypic FAM 161A interactions, as well as heterotypic interactions with paralog FAM161B (family with sequence similarity 161, member B) [33] . FAM 161B interacts with TACC3 (transforming, acidic coiled-coil containing protein 3), which in turn is involved in centrosome-dependent microtubule assembly, kinetochore attachment, chromosome alignment and mitotic exit [36] . FAM 161A could therefore be involved in maintenance of the microtubule axoneme along the connecting cilium or protein transport between the inner segment (IS) and outer segment (OS) [32,33] .

FAM161Ar i758-i5_i758-i6ins238 affects splicing of exon 5 of the gene, resulting in a truncated protein, including the loss of approximately 44 amino acids of the UPF0564 conserved domain. However, Bandah-Rozenfeld et a/, reported that the N-terminus of the UPF0564 domain is sufficient for homotypic and heterotypic interaction with FAM161B [18] . The truncated protein product is therefore expected to be functional in this regard. As the discovery of FAM161A involvement in retinal disease was relatively recent, very little is known about the protein's structure and function of the protein in visual and other pathways. Further investigations are required to elucidate the precise pathways in which FAM 161A is involved, which may lead to the identification of novel functional domains in the C-terminus of the protein. To this end, the canine model described here could be particularly useful, as no other animal models have been reported.

The presence of FAM161A mutant mRNA transcripts in the blood of an affected dog implies that the truncated transcript is not subjected to nonsense-mediated decay. A truncated protein may therefore be expressed, although this would need to be confirmed by comparing FAM 161A protein levels in FAM161A^~f~ dogs with protein levels in FAM161A^+/+ dogs. If this is the case, the truncated protein product must be sufficient to cause retinal degeneration.

PRA caused by the variant described here has an average age at diagnosis of 4.89 years and this is indicative of a late age of onset and consistent with observations in human patients in which the age of onset was in the 2^nd or 3^rd decade [19] . Given that FAM 161A is expressed in multiple tissues, it would be interesting to determine whether a more severe change to the protein, such as a knock-out, would result in a more severe retinal or even systemic phenotype.

The discordant TS PRA cases i. e. FAM161A^+/+ and FAM161A^+/~ tended to develop PRA at a later age, with an average age at diagnosis of 7.01 years, which is consistent with the segregation of a second form of PRA in the TS breed.

PRA in the TS has not previously been associated with any genetic variants. Using a GWA mapping approach, a novel candidate variant, FAM161A_c.₁₇₅₈-i5_i758- i6ins238, was identified that is likely to represent a major causal mutation for PRA in the TS. While this mutation does not account for all cases of PRA in this study, suggesting that there are additional loci causing PRA in this breed, it does appear to be highly penetrant and a major cause of PRA in this breed. While PRA3 is also present in TTs, as they are closely related and the mutation has not been found in any other breeds, the mutation appears to be confined to these two breeds.

REFERENCES

1. Parry HB (1953) Br J Ophthalmol 37 : 487-502.

2. Petersen-Jones S (2005) J Small Anim Pract 46 : 371-380.

3. Andre C, et al. (2008) Pratique Medicale et Chirurgicale de I'Animal de Compagnie 43 : 75-84.

4. Grondahl J (1987) Clin Genet 31 : 255-264.

5. Haim M, Holm NV, Rosenberg T (1992) Acta Ophthalmol (Copenh) 70 : 178-186.

6. Pagon RA (1988) Surv Ophthalmol 33 : 137- 177.

7. Daiger SP, Bowne SJ, Sullivan LS (2007) Arch Ophthalmol 125 : 151-158. 8. Grail A, et al. (2012) Nat Genet 44: 140-147.

9. Beggs AH, et al . (2010) Proc Natl Acad Sci U S A 107 : 14697-14702.

10. Sidjanin DJ, et al . (2002) Hum Mol Genet 11 : 1823-1833.

11. Mellersh CS, et al. (2006) Genomics 88 : 293-301.

12. Howell JM, et al. (1997) Neuromuscul Disord 7 : 325-328.

13. Acland GM, et al . (2001) Nat Genet 28 : 92-95.

14. Mount JD, et al. (2002) Blood 99 : 2670-2676.

15. Ponder KP, et al. (2002) Proc Natl Acad Sci U S A 99 : 13102-13107.

16. Bainbridge JW, et al. (2008) N Engl J Med 358 : 2231-2239.

17. Beltran WA, et al. (2012) Proc Natl Acad Sci U S A 109 : 2132-2137.

18. Bandah-Rozenfeld D, et al . (2010) Am J Hum Genet 87 : 382-391.

19. Langmann T, et al . (2010) Am J Hum Genet 87 : 376-381.

20. Minnick MF, Stillwell LC, Heineman JM, Stiegler GL (1992) Gene 110 : 235- 238.

21. Johnson JM, et al. (2003) Science 302 : 2141-2144.

22. Lewis BP, Green RE, Brenner SE (2003) Proc Natl Acad Sci U S A 100 : 189-192.

23. Fairbanks DJ, Anderson WR, editors (1999) Genetics: The Continuity of Life : Brooks/Cole Publishing Company.

24. Reed R, Maniatis T (1988) Genes Dev 2 : 1268-1276. 25. Reed R, Maniatis T (1985) Cell 41 : 95-105.

26. Lin L, et al . (1999) Cell 98 : 365-376.

27. Clark LA, Wahl JM, Rees CA, Murphy KE (2006) Proc Natl Acad Sci U S A 103 : 1376-1381.

28. Roy-Engel AM, et al. (2002) Genome Res 12 : 1333-1344.

29. Miyadera K, Brierley I, Aguirre-Hernandez J, Mellersh CS, Sargan DR (2012) PLoS ON E 7 : e51598.

30. Downs LM, et al. (2012) Animal Genetics 44: 169-177.

31. Downs LM, et al. (2011) PLoS One 6 : e21452.

32. Di Gioia SA, et al . (2012) Hum Mol Genet.

33. Zach F, et al . (2012) Hum Mol Genet.

34. Freund CL, et al . (1997) Cell 91 : 543-553.

35. den Hollander Al, et al. (2007) Nat Genet 39 : 889-895.

36. Gomez-Baldo L, et al . (2010) Cell Cycle 9 : 1143-1155.

37. Purcell S, et al. (2007) Am J Hum Genet 81 : 559-575.

38. Rozen S, Skaletsky H (2000) Methods Mol Biol 132 : 365-386.

39. Agilent Technologies eArray tool.

40. Li H, Durbin R (2009) Bioinformatics 25 : 1754-1760.

41. McKenna A, et al . (2010) Genome Res 20 : 1297-1303.

42. Thorvaldsdottir H, Robinson JT, Mesirov JP (2012) Brief Bioinform.

43. Thompson JD, Higgins DG, Gibson TJ (1994) Nucleic Acids Res 22 : 4673- 4680.

44. Bonfield JK, Smith K, Staden R (1995) Nucleic Acids Res 23 : 4992-4999.

Claims

1. Use of the FAM161A gene as a biomarker for the in vitro prognosis of progressive retinal atrophy developing in a canine mammal.

2. An in vitro method of prognosing progressive retinal atrophy in a canine mammal, the method comprising the step of detecting genetic variation within the FAM161A gene.

3. The use or method as defined in claim 1 or claim 2, wherein the canine mammal is a dog which is a breed selected from Tibetan Spaniel and Tibetan Terrier.

4. The use or method as defined in claim 3, wherein the canine mammal is a Tibetan Spaniel.

5. The method as defined in any of claims 2 to 4, which comprises the steps of:

(i) providing a sample of nucleic acid from the canine mammal;

(ii) detecting genetic variation within the FAM161A gene in the nucleic acid sample; and

(iii) correlating the result from step (ii) with the prognosis of progressive retinal atrophy developing in the canine mammal .

6. The method as defined in claim 5, wherein the nucleic acid is genomic DNA.

7. The method as defined in any one of claims 2 to 6, wherein the genetic variation is within intron 4 of the FAM161A gene.

8. The method as defined in any one of claims 2 to 7, wherein the genetic variation comprises an insertion mutation within the FAM161A gene.

9. The method as defined in claim 8, wherein the insertion mutation comprises a SINE insertion at position 64,974,130 on chromosome 10.

10. The method as defined in any one of claims 2 to 5, wherein detection of the genetic variation is performed by detecting the absence of exon 5 in mRNA of the FAM161A gene.

11. The method as defined in any one of claims 2 to 10, which additionally comprises the step of establishing whether or not the canine mammal is heterozygous or homozygous for the genetic variation within the FAM161A gene.

12. The method as defined in claim 11, wherein if the canine mammal is homozygous for the genetic variation within the FAM161A gene, it is prognosed as a canine mammal that will suffer from progressive retinal atrophy.

13. The method as defined in claim 11, wherein if the canine mammal is heterozygous for the genetic variation within the FAM161A gene, it is selected as being suitable for breeding with a canine mammal of the same breed which is homozygous for the wild-type FAM161A gene.

14. The method as defined in claim 11, wherein if the canine mammal is homozygous for the wild-type FAM161A gene, it is selected as being suitable for breeding with a canine mammal of the same breed which is homozygous or heterozygous for the wild-type FAM161A gene.

15. The method as defined in any one of claims 2 to 14, wherein the sample is assessed for one or more other markers which are linked or associated with canine disorders.

16. The method as defined in any one of claims 2 to 15, wherein the detection step is performed by determining the binding of oligonucleotide probes to the nucleic acid sample, wherein the probes comprise all or part of the wild-type or mutant FAM161A gene.

17. The method of claim 16, wherein the oligonucleotide probes bind within intron 4 of the FAM161A gene.

18. The method as defined in claim 17, wherein the oligonucleotide probes are :

Forward mutant: 5 '- G G ATC C CTTTATTTG ATTTTA G A A AG - 3 ' (SEQ ID NO : 1); Forward wild-type: 5'-TCCCTTCCTTTTATTTGATTTTAGAAAG-3' (SEQ ID NO : 2); and

Reverse: 5 '- 6 FAM - CAACAAACACAACCTG AGCAA- 3 ' (SEQ ID NO : 3).

19. The method as defined in any one of claims 2 to 18, wherein the detection step is performed by amplifying all or part of the FAM161A gene.

20. The method as defined in claim 19, wherein the amplified region is less than 500 nucleotides in length, such as less than 400 nucleotides in length, in particular 200 to 400 nucleotides in length.

21. Oligonucleotide probes for use in a method of prognosing progressive retinal atrophy in a canine mammal, wherein said oligonucleotide probes detect a genetic variation within the FAM161A gene and are as defined in claim 18.

22. A kit for use in a method of prognosing progressive retinal atrophy in a canine mammal, wherein said kit comprises:

(a) oligonucleotide probes which detect a genetic variation within the FAM161A gene and are as defined in claim 18; and

(b) means for providing a test sample from the canine mammal .

23. A method of treating progressive retinal atrophy in a canine mammal, which comprises assessing the progressive retinal atrophy status of a canine mammal by use of a method as defined in any of claims 2 to 20 and if the canine mammal is identified as affected by progressive retinal atrophy, treating said canine mammal to prevent or reduce the onset of progressive retinal atrophy.

24. A method of treating progressive retinal atrophy in a canine mammal, which comprises increasing the level of non-mutant, wild-type FAM161A gene expression and/or FAM161A gene product activity in the canine mammal.