Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

A Novel Variant in CMAH Is Associated with Blood Type AB in Ragdoll Cats

  • Barbara Gandolfi ,

    Affiliation Department of Veterinary Medicine and Surgery, College of Veterinary Medicine, University of Missouri—Columbia, Columbia, Missouri, 65211, United States of America

  • Robert A. Grahn,

    Affiliation Veterinary Genetics Laboratory, School of Veterinary Medicine, University of California Davis, Davis, California, 95616, United States of America

  • Nicholas A. Gustafson,

    Affiliation Department of Veterinary Medicine and Surgery, College of Veterinary Medicine, University of Missouri—Columbia, Columbia, Missouri, 65211, United States of America

  • Daniela Proverbio,

    Affiliation Veterinary Transfusion Unit (REV), Department of Health, Animal Science and Food Safety (VESPA), University of Milan, 20133, Milan, Italy

  • Eva Spada,

    Affiliation Veterinary Transfusion Unit (REV), Department of Health, Animal Science and Food Safety (VESPA), University of Milan, 20133, Milan, Italy

  • Badri Adhikari,

    Affiliation Department of Computer Science, Informatics Institute, University of Missouri–Columbia, Columbia, Missouri, 65211, United States of America

  • Janling Cheng,

    Affiliation Department of Computer Science, Informatics Institute, University of Missouri–Columbia, Columbia, Missouri, 65211, United States of America

  • Gordon Andrews,

    Affiliation Department of Diagnostic Service/Pathobiology, College of Veterinary Medicine, Kansas State University, Manhattan, Kansas, 66506, United States of America

  • Leslie A. Lyons,

    Affiliation Department of Veterinary Medicine and Surgery, College of Veterinary Medicine, University of Missouri—Columbia, Columbia, Missouri, 65211, United States of America

  • Chris R. Helps

    Affiliation Langford Veterinary Services, University of Bristol, Bristol, BS40 5DU, United Kingdom


13 Mar 2018: Gandolfi B, Grahn RA, Gustafson NA, Proverbio D, Spada E, et al. (2018) Correction: A Novel Variant in CMAH Is Associated with Blood Type AB in Ragdoll Cats. PLOS ONE 13(3): e0194471. View correction


The enzyme cytidine monophospho-N-acetylneuraminic acid hydroxylase is associated with the production of sialic acids on cat red blood cells. The cat has one major blood group with three serotypes; the most common blood type A being dominant to type B. A third rare blood type is known as AB and has an unclear mode of inheritance. Cat blood type antigens are defined, with N-glycolylneuraminic acid being associated with type A and N-acetylneuraminic acid with type B. Blood type AB is serologically characterized by agglutination using typing reagents directed against both A and B epitopes. While a genetic characterization of blood type B has been achieved, the rare type AB serotype remains genetically uncharacterized. A genome-wide association study in Ragdoll cats (22 cases and 15 controls) detected a significant association between blood type AB and SNPs on cat chromosome B2, with the most highly associated SNP being at position 4,487,432 near the candidate gene cytidine monophospho-N-acetylneuraminic acid hydroxylase. A novel variant, c.364C>T, was identified that is highly associated with blood type AB in Ragdoll cats and, to a lesser degree, with type AB in random bred cats. The newly identified variant is probably linked with blood type AB in Ragdoll cats, and is associated with the expression of both antigens (N-glycolylneuraminic acid and N-acetylneuraminic acid) on the red blood cell membrane. Other variants, not identified by this work, are likely to be associated with blood type AB in other breeds of cat.


Cytidine monophospho-N-acetylneuraminic acid hydroxylase (CMAH) is associated with the production of the sialic acids present on cat erythrocytes [1]. CMAH converts N-acetylneuraminic acid (NeuAc) to N-glycolylneuraminic acid (NeuGc) by substituting one of the hydrogen atoms in the methyl moiety with a carboxyl group [2,3]. Feline CMAH was the first gene shown to control blood types in non-primate mammals [4]. The gene is non-functional in humans [5,6], New World monkeys [7], birds and reptiles [8]. The ability to synthesize NeuGc was lost in hominids ~3 million years ago when an Alu element disrupted exon 6 of CMAH [5], and ~30 million years ago in the New World monkeys, through an inversion of exons 4 to 13 [7]. Loss of function of genes involved in membrane glycan synthesis is often an evolutionary defense mechanism to combat pathogen cell invasion [9,10], such as in the case of the malaria parasite Plasmodium falciparum, which prefers to bind to NeuGc [11,12]. By inactivating CMAH, the NeuAc precursor accumulates and NeuGc is absent on the erythrocyte cell surface, thus reducing parasite infection in humans and promoting a positive selection during evolution, which results in the enrichment of “resistant” blood type antigens.

The cat has one major blood group with three serotypes; the most common being blood type A that is dominant to type B. A third rare blood type is known as AB [13], and from breeding experiments has an unclear mode of inheritance due to apparently different inheritance patterns in different breeds [14]. Cat blood type antigens are defined, NeuGc is associated with type A and NeuAc with type B. Blood type AB is characterized by agglutination with antibodies against both NeuGc and NeuAc [1]. In a previous study, Bighignoli et al. (2007) demonstrated that the genomic structure of cat CMAH was similar to the homologous gene in other species [4]. The cat AB blood group system resulted from variant(s) in CMAH that prevented the conversion of NeuAc to NeuGc, hence type B to type A. Several variants, including SNPs upstream of the start codon, a deletion in the 5’ UTR of exon 1, and three variants in the coding region were concordant with A and B blood types in 18 different breeds of cat. Moreover, in the same study, the cats that tested positive for both NeuGc and NeuAc (type AB) could not be genetically differentiated from blood type A cats [4], hence blood type AB remained genetically uncharacterized. Note that the numbering in the original Bighignoli et al. paper is mis-aligned by 3 bases compared to the position in the feline coding sequence (CDS). Positions are updated in the current work, hence the original c.139G>A variant is actually at position c.142 of the CMAH CDS.

The frequencies of the different blood types vary among cat breeds and random bred populations with different origins. Several studies of cats from different continents confirmed blood type A to be the most common with a prevalence from 73% in Australia [13], 88.2% in China [15] to 95.3% in the United States West coast and 99.7% in United States Northeast [16]. Blood type B varies from 0.3% to 33% [1619] and type AB is the rarest (0.2% to 20%) [13,20,21]. Blood type AB is common in Ragdolls with a frequency of 20% [22], Devon Rex (1.4%), British shorthair (1.6%) and particular random bred populations. In Japan, the proportion of blood type AB in the random bred population is 9.7% [23], 6.8% in Portugal [24], 5% in the UK [20], 2.3% in Brazil [25], 1.6% in Australia [26], and less than 0.2% in Canada [27] and the USA [14]. Whilst founder effects can explain the variation in frequency between breeds, and are probably the cause of the high percentage of type AB in Ragdolls, the variation between random bred populations strongly suggests an evolutionary enrichment of one blood type over another.

Cats possess naturally occurring antibodies against the blood antigen they lack [13,28,29]. Blood type B cats have strong anti-A alloantibodies [13,28] and blood type A cats have weak anti-B alloantibodies [28], while blood type AB cats have no alloantibodies since they express both antigens. Adverse reactions (such as vomiting, pyrexia, hemolysis, electrolyte disturbances, circulatory overload and urticarial) are observed when transfusing type A blood to a B recipient. A mild transfusion reaction can be observed if a type A recipient receives type B blood. However, if the type A cat has been previously exposed to type B blood, such as during pregnancy, serious transfusion reactions can occur, although these tend to be less severe and are less common than when type A blood is given to a type B cat. [16,29]. Blood type AB cats may develop a hemolytic reaction after a type B blood transfusion (due to strong anti-A alloantibodies in the blood donor) [14], hence transfusing type AB or type A blood is recommended. The lack of a feline universal blood donor requires that all cats must be blood typed and cross-matched before any transfusion and/or transplant. The characterization of blood type AB, and subsequent development of a genetic test, will increase the precision of genetic blood typing and help breeding programs. Moreover, once the type AB genetic test is developed, a substantial screening of blood types can be easily conducted and will allow the prevalence in different breeds and populations to be determined, offering insights into natural selection in different environmental conditions and resistance to pathogens associated with certain blood types. This study presents the genetic investigation of feline blood type AB in Ragdoll cats and confirms the power of the feline SNP array in detecting a phenotype–locus association. A genome-wide association study showed the blood type AB locus to be in close proximity to CMAH, which has previously been shown to control blood types A and B. Detailed investigation and screening of hundreds of cats supports the characterization of a novel feline blood type variant in Ragdoll cats.


Genome-wide association study

The datasets supporting the results are included within additional supplementary files (including S1 and S2 Datasets).

The Ragdoll cases were selected by pedigree analysis to have minimal relationship to the controls, and 37 cats, including 22 cases (blood type AB) and 15 controls (blood type A, Ab or B), were submitted for SNP array genotyping (S1 Dataset, S2 Dataset). Testing for stratification revealed that four cats (one control and three cases) had a P̂ > 0.35, suggesting familiar relationships within the sample cohort. These samples were excluded from the genome-wide association study to reduce inflation of the results, leaving 19 cases and 14 controls for analysis. After the exclusion of the related samples, the genomic inflation (λ) was reduced from 1.72 to 1.50. Testing for stratification by multi-dimensional scaling (MDS) revealed that the cases and controls formed one main overlapping cluster, thus no additional samples were excluded (Fig 1A).

Fig 1. Multidimensional scaling and Manhattan plot summarizing the GWAS for Ragdoll cats with blood type AB.

a. MDS showing the distribution of the Ragdoll cases and controls included in the analyses. Significant overlap of cases and controls suggests minimal substructuring in the dataset. b. The plot represents the Praw (top) and Pgenome (bottom) values of the SNPs included in the GWAS case–control analysis. The black dashed line (bottom) suggests the SNPs with genome-wide significance after permutation. Two SNPs on chromosome B2 (positions 4,487,432 and 4,449,822, respectively) remained significantly associated with the AB blood type phenotype.

After initial quality control of the 62,897 SNPs on the array, 47,239 markers were included in the analysis; 15,557 were excluded for a minor allele frequency < 0.05 and 101 SNPs failed genotype frequency > 20% across all the samples. By analyzing 19 cases and 14 controls, a significant association was detected with SNPs on cat chromosome B2 (Fig 1B), and correcting Praw values for multiple hypotheses testing by permutation, two adjacent SNPs on chromosome B2 at positions 4,487,432 and 4,449,822 respectively (S1 Table), were genome-wide significant (Pgenome = 0.00001 and Pgenome = 0.004) and remained significant after Bonferroni correction (1.8e-6 and 0.0097). The two SNPs are in linkage disequilibrium with the phenotype, the most common haplotype having a frequency of 79% in the cases and 10% in the controls, defining a short (~ 40 Kb) haplotype block. The haplotype block is in the intergenic region between two genes: the leucine-rich repeat-containing protein 16A isoform 2 (LRRC16A) located ~50 Kb upstream of the haplotype block and CMAH, which is located ~40 Kb downstream. Since CMAH has been shown to control feline blood types A and B, this gene was further examined for cDNA and genomic variants that might be responsible for blood type AB.

CMAH cDNA and genomic DNA comparison

Sequence data for CMAH cDNA was derived from six cats: four Ragdoll blood type AB, one random bred blood type AB, one random bred type B (serology B, genotype bb) and compared to the publicly available random bred type A (serology A, genotype AA) cDNA sequence (EF127684.1). Sequence analyses of the coding region identified nine cDNA variants, including five missense mutations. Four of the nine variants had previously been identified by Bighignoli et al. (2007), thus five CMAH variants (c.327A>C, c.364C>T, c.1158T>C, c.1269A>G, c.1398G>T) were novel to this study (Table 1, S1 Fig). Note that in Table 1, SNP positions within the CMAH CDS have been updated. Only the c.364C>T variant in exon 4, predicted in silico to change a nonpolar and hydrophobic amino acid (proline) to a polar and uncharged residue (serine) at codon 122, was present in the 5 type AB cats and absent in the type A and B cats. The c.364C>T variant was further investigated by genomic analysis in 280 cats with known blood type (see below).

Table 1. Comparison of CMAH cDNA variants in cats with blood types A, B and AB.

The c.327A>C variant was unique to the blood type B random bred cat. Two of the three synonymous mutations (c.1158T>C, c.1269A>G) were found exclusively in the blood type A random bred (from reference sequence EF127684.1), while c.1398G>T was found in the type A and B random bred cats and was not present in the type AB phenotype cats (Table 1).

CMAH variant genotyping

The newly identified c.364C>T variant in exon 4 was genotyped in 280 cats with known blood types defined by serology and was found in 89 of 115 type AB cats with at least one allele coding for blood type A. The variant was only detected in Ragdoll and random bred cats. Four type AB Ragdoll and 13 type AB random bred cats were wild-type for c.364C>T. The variant was not associated with blood type AB in Devon rex (n = 8) and British shorthair (n = 1) cats and was never found in blood type A or B cats (n = 164) (Table 2).

Table 2. Domestic cats with known blood type serology genotyped for the c.364C>T variant in CMAH exon 4.

Other CMAH variants were also investigated in the 280 cats: the 18 bp deletion within the CMAH 5’UTR, the exon 2 c.142G>A and the exon 2 c.139C>T (LVS cats only) (S2 Table). The exon 2 c.142G>A and c.139C>T variants were originally numbered c.139G>A and c.136C>T in Bighignoli et al. (2007) and Tasker et al. (2014), respectively. The 18 bp deletion and c.142G>A variant, previously associated with feline blood type B, were found on the feline B allele, confirming the findings from the Bighignoli et al. (2007) study. However, nine cats with known blood type A or B serology showed a discrepancy between the deletion and the c.142G>A variant. In seven cats, the c.142G>A was associated with the blood type confirmed by serology while the deletion was discordant. For two cats, both variants were discordant and suggested a serology mistype. For one cat, the deletion was concordant with the blood type, suggesting the missense mutation (c.142G>A) as more likely causal for blood type B, although an error in blood serotyping cannot be excluded.

To determine the allele frequency of the c.364C>T variant, 786 cats from many breeds with unknown blood type were genotyped for the newly identified variant, as well as the 5’ UTR deletion and the c.142G>A missense mutation. The c.364C>T variant was only detected in the Ragdoll, random bred and Cornish rex cats; the frequency of the T allele in these breeds was 8.8%, 10% and 2.4%, respectively (Table 3). Cats with the c.364C>T variant were genotyped as AB/AB, AB/A or AB/b. In the 786 cats tested, 53 (6.7%) showed a discrepancy between the 5’ UTR deletion and the c.142G>A variant (S4 Table).

Table 3. Frequency of the CMAH c.364C>T polymorphism in different cat breeds with unknown serology.

Since the c.364C>T variant was not concordant in all blood type AB cats, genomic analysis of CMAH was conducted on two type AB Devon rex, one type AB British shorthair, one type AB random bred cat and one type A random bred cat that were wild-type for the c.364C>T variant. Twelve variants were detected: seven in the coding region, one in the 5’ UTR and four in intronic regions. Five novel variants were identified, four located in introns (I3 c.307+22A>G, I9 c.1112+29T>G, I9 c.1112+66G>A, I9 c.1113-70G>A) and one silent mutation in the coding region, (c.884G>A). Only the variant localized 70 bp upstream of exon 10 (I9 c.1113-70G>A), was concordant with blood type AB, but was not predicted to disrupt any conserved splice site regions and was not investigated in detail (S3 Table).

Computer modeling of feline CMAH protein structure and c.142G>A (V48M)

Templates with known tertiary structure were identified for two regions of the CMAH protein sequence (from residues 21 to 94 and from 207 to 294) using PSI-BLAST [30] with an e-value threshold of 0.005. For the first region (residues 21 to 94), the top template is a Rieske iron-sulfur protein 2PQZ, which covers the first variant site of interest (c.142G>A, V48M). Iron-sulfur [Fe-S] clusters are ubiquitous and required to sustain central life processes. Fe-S clusters participate in electron transfer, substrate binding/activation, gene expression regulation and enzyme activity [31]. The template (S2 Fig) shows that the metal-ion binding site is spatially distant from the V48M variant. The other template covering the second region of the CMAH sequence (residues 207 to 294, S2 Fig) is part of the complex of a large ribosomal subunit from Deinococus radiodurans (3JQ4) with a zinc binding site. The 3-D protein structure surrounding the novel P122S variant could not be modeled due to the lack of an available 3-D template.


Sialic acids are a family of sugars that share a nine carbon backbone and are typically found on cell surfaces and glycan molecules. About 40 sialic acids are documented with the most common being NeuAc, which is a substrate for the synthesis of NeuGc by CMAH. These two sialic acids are the major determinants of feline blood types A and B. A third, rare and genetically uncharacterized blood type (AB) also exists with differing frequencies amongst cat populations. The A/B blood group system is clinically important in cats because mismatched breeding and transfusions can cause life threatening hemolytic reactions without prior sensitization [14]. Therefore, a complete genetic characterization of all feline blood types is crucial so that genetic testing can accurately report an individual’s blood type.

Recent advances in feline genomics, such as a new genome assembly [32] and the availability of a high density SNP array chip, have accelerated the discovery of several genetic diseases [3335] and phenotypic traits [36,37] in the domestic cat. The Ragdoll is a relatively new breed with western origins (California, USA) [38,39] and characterized by point coloration and long hair [40,41]. Ragdolls are not significantly related to any other breed. In a previous study, Alhaddad et al. (2013) [42] reported that a recently developed breed might have longer haplotype blocks, enabling genome-wide association studies with fewer individuals [36,37]. Strong selection for fewer aesthetic traits and a breeding program dependent on very few founders results in high levels of LD and long haplotype blocks, characteristic of identity by descent traits. In this study, using only 33 individuals, an association of blood type AB with a variant on cat chromosome B2 in the Ragdoll breed was detected. Likely, during the Ragdoll breed development, a cat with the AB phenotype was extensively used in the breeding program, introducing the variant into the breed with a high frequency.

Due to the overall low prevalence of blood type AB, it was unclear whether CMAH was the locus associated with the AB blood type phenotype, or whether there were different loci involved, in fact, the possibility of cis-acting alleles was previously suggested [13]. Griot-Wenk et al. [14] presented strong evidence that type AB was allelic, but the presence of a third allele could not explain blood type inheritance in British shorthair and Somali cats, additionally, the inheritance of blood type AB was never studied in Ragdoll cats, hence the need for an association study. The GWAS indicated that the locus influencing the rare blood type AB resided on cat chromosome B2, within one of two genes: LRRC16A and CMAH. Since LRRC16A is responsible for the inhibition of actin filament capping, which results in enhancing actin polymerization [43], CMAH remained the top candidate gene for the blood type AB phenotype, given that the gene is the major locus controlling blood types A and B in cats [4]. Although Bighignoli et al. [7] included a type AB Siamese cat in their study they were not able to identify a causal variant for blood type AB. Furthermore, Ragdoll cats were not included in the previous study and their CMAH CDS was only determined in the current study. In this study, a novel variant in CMAH exon 4 (c.364C>T) that causes a proline to serine amino acid substitution at position 122 (p.122P>S) was identified. Proline is the only amino acid where the side chain is connected to the protein backbone twice and is often found in very tight turns in protein structures. The proline side chain is non-reactive and is rarely involved in a protein’s active or binding sites, while serine is quite common in protein functional centers and is fairly reactive. Prolines are considered important in protein structure, hence this substitution might be associated with an abnormal protein conformation, resulting in a partial or aberrant enzyme function. The newly identified variant is highly associated with blood type AB in Ragdoll cats, with only four of the 58 blood type AB Ragdoll cats being homozygous for the wild-type exon 4 allele. In these discordant cats, another variant at a second locus or at the same locus in the non-coding sequence may be the cause of their blood serotype or it could be due to a database or sampling error. Alternatively, their blood phenotype could be incorrect due to inaccurate serotyping or a change in blood type serology due to sickness. A previous study showed that humans absorb a portion of ingested NeuGc and incorporate a small amount into newly synthesized glycoconjugates [44,45]. Julien et al. [46] and Mortezai et al. [47] demonstrated that the presence of NeuGc and NeuAc improves local inflammatory responses, hence causing a temporary variation in blood serology results.

Feline blood typing can be commercially performed using several methods including a card that uses monoclonal antibodies [48], a matrix gel column containing serum or lectin [49], a slide and tube assay that uses lectin and anti-A sera [50] and a commonly used typing kit based on immunochromatographic diffusion of red blood cells [51], which has recently been shown to outperform the card method with 100% specificity and sensitivity in anemic and non-anemic cats [52]. Seth et al. (2011) [51] compared feline blood typing methods and showed that agreement between the methods varied from 99.4% to 91.4% when compared to the tube assay. Therefore, potentially the four discordant Ragdoll cats submitted as blood type AB, were actually type A or B [51]. In the domestic cat population, the newly identified variant was found in 73% of blood type AB cats, and was absent in type AB cats from the British shorthair and Devon rex breeds. Direct sequencing of the CMAH CDS in Devon rex and British shorthair cats with blood type AB did not suggest any previously undetected missense or splice site disrupting variants, suggesting blood type AB is controlled by a different locus, or a variant in the same locus in a non-coding region, in these breeds.

Both NeuGc and NeuAc sialic acids appear to be present at low levels on the red cell surface of blood type AB cats [1]. The presence of both antigens on type AB cats’ erythrocytes can be explained by a reduction in CMAH activity. Several studies showed that the amount of NeuGc incorporated on the cell membrane is regulated by the level of hydroxylase activity [53,54], hence the newly identified c.364C>T variant may reduce CMAH activity. Tasker et al. [20] reported a previously unknown c.136C>T (numbered c.139C>T herein) variant associated with feline blood type B; potentially every novel variant that knocks-out CMAH activity could be associated with blood type B. By comparison, all variants that reduce CMAH activity may be associated with blood type AB, suggesting the possibility of more, as yet unidentified, variants associated with feline blood types and the need for further molecular investigation. A three-dimensional prediction of the effect of the c.142G>A (V48M) missense mutation associated with blood type B suggests that CMAH is still capable of binding the substrate (since the substitution does not disrupt the [Fe-S] domain) but, since NeuGc is not synthesized, the enzyme activity must be compromised by the variant [1]. Unfortunately, no insights were gained from the protein structure prediction on the effects of the substitution associated with blood type AB, mainly due to the lack of a viable protein template for the region.

The AB serotype is allelic to A and B in cats, and the variant associated with type AB resides on the enzymatically active type A allelic background, and is likely associated with a reduction of CMAH activity and the presence of both sialic acids on the red cell surface. A three allele system is proposed: A > aAB > b, where a type A cat may have AA, AaAB or Ab genotype. The type B cats are always homozygous bb. Possible genotypes / phenotypes (serotypes) would be AA (type A); AaAB (type A); Ab (type A); aABb (type AB); aABaAB (type AB), and bb (type B). Whilst the majority of blood type AB cats in the study were aABb or aABaAB a significant number (n = 27) were genotype AaAB, and would be predicted to be blood type A. This apparent discordancy could be explained by a second, yet to be identified, type AB variant on the A allele, and is not an unexpected finding since other breeds with blood type AB were shown not to have the c.364C>T variant. An in-depth molecular investigation of samples with discrepancies between genotype (c.142G>A, c.364C>T) and blood phenotype is needed in future studies to detect other alleles involved in feline blood types.


This study identified a variant associated with the rare feline AB blood type in Ragdoll cats. A genome-wide association study suggested CMAH as a candidate gene and a newly identified c.364C>T missense mutation is probably linked with the feline AB blood type. Moreover, the variant is likely associated with a reduction of CMAH enzyme activity and the resultant expression of both antigens (NeuGc and NeuAc) on the red cell membrane. Since blood type AB cats from other breeds did not have the Ragdoll type AB variant, at least one other variant must exist that also results in a type AB phenotype.


Ethics statement

This study was approved by the Animal Care and Use Committee (ACUC) of the University of Davis (protocol # 16991) and the University of Missouri (protocol # 7808). Italian samples were collected with the informed consent of owners during routine wellness visits. Several samples were acquired by specialist in the field, such as a veterinarian, or voluntarily donated by owners and breeders with consent to use the specimens for research purposes.

Sample collection

European and USA cat owners and breeders voluntarily donated EDTA anti-coagulated whole blood or buccal swabs from blood type A, B and AB cats (n = 134). Additionally, 138 DNA samples from cats blood typed by hemagglutination and alloantibody assays [13] were also provided by several laboratory services: (1) Oy Triniini Company, Helsinki, Finland, (2) Langford Veterinary Services (LVS), Bristol, UK, (3) Department of Diagnostic Medicine/Pathobiology, College of Veterinary Medicine, Kansas State University (4) and the University of Milan, Italy. The complete sample set (N = 280) included 110 type A samples, 54 type B and 116 type AB. The sample set included 73 random bred cats and 15 breeds (87 Ragdoll, 40 Birman, 40 Devon Rex, 18 British shorthair and 22 cats from 11 other breeds, Table 2).

The Veterinary Genetic Laboratory (VGL) at the University of California, Davis, genotyped 786 samples with unknown blood serology/phenotypes, from several different breeds and random bred cats (S4 Table).

DNA from 37 Ragdolls (22 type AB and 15 type A, Ab or B) were submitted to Geneseek Inc. (Lincoln, NE) for genotyping analysis on the array. DNA extraction and quality control was conducted as previously described [33]. Supplementary files contain the genotyping data (S1 Dataset) and the location of each SNP typed on the array (S2 Dataset). Blood from four Ragdoll type AB, one random bred type AB and a random bred type B was also collected in PAXgeneTM (Qiagen, Valencia, CA) tubes for RNA extraction.

Availability of Data and Materials

The supporting file section contains the raw genotyping data for the 37 Ragdolls included in this study (S1 Dataset) and the genomic location of each variant (S2 Dataset) typed on the Illumina Infinium Feline 63K iSelect DNA array (Illumina, Inc., San Diego, CA).

Case-control genome wide association analysis

Initially, genomic DNA from 22 type AB Ragdolls (cases) and 15 type A, Ab or B Ragdolls (controls) was genotyped on the Illumina Infinium Feline 63K iSelect DNA array. SNP genotype data analysis was performed with PLINK [55] and subjected to standard quality control. SNPs were excluded for poor genotyping rates (< 80%), low minor allele frequencies (< 0.05%) and individuals were excluded for low genotyping success (< 80%). Inflation of p-values was evaluated by calculating the genomic inflation factor (λ). To reduce genomic inflation, individuals with a proportion of IBD > 0.35 were excluded from the downstream analyses. A classic multi-dimensional scaling (MDS) with 2-dimensions was performed in PLINK to evaluate population substructure within cases and controls. The case-control association analysis was performed and corrected with 100,000 t-max permutations (Fig 1). T-max calculated p-values were considered genome-wide significant at p < 0.05. After each correction, the P value of the highly associated SNPs was evaluated and reported in S1 Table. All genomic positions are reported in accordance with the Felis_catus 6.2/felCat5 build of the feline genome [32].

CMAH gene and CDS analysis

Total RNA was extracted using the PAXgene™ kit (Qiagen) from whole blood from six cats including four Ragdoll type AB, one random bred type AB and one random bred type B. Complementary DNA was synthesized using SuperScript III (Invitrogen, Carlsband, CA) by reverse transcription of 1 μg of total RNA with gene specific primers (CMAH-1R, CMAH-2R, CMAH-3R, CMAH-4R) and oligo dT to obtain UTRs and CDS (S5 Table). Each cDNA sample served as template for PCR using 1 μM primers combined as follow: CMAH-1F + CMAH-1R, CMAH-2F + CMAH-2R, CMAH-3F + CMAH-3R, CMAH-4F + CMAH-4R, CMAH-3UTR+ polyT. The PCR conditions were: 1.5 mM MgCl2, 1X Buffer, 0.1 U of AmpliTaq Gold DNA polymerase (Thermo Fisher Scientific, Waltham, MA) and 2 μl of cDNA in a total volume of 20 μl. The PCR was: initial denaturation at 94°C for 4 min followed by 40 cycles at: 94°C for 30 sec, 62°C for 30 sec, 72°C for 1 min and a final extension at 72°C for 20 min. The PCR products were purified and sequenced as previously described [4]. The 5’ UTR was obtained as previously described using the gene specific primer CMAH-5UTR (S5 Table), the PCR conditions included an initial denaturation of 5 min followed by 40 cycles of 30 sec denaturation at 94°C, annealing for 30 sec at 56°C and extension at 72°C for 30 sec, followed by 10 min extension at 72°C.

Genomic analysis of CMAH was conducted on four type AB cats (one random bred, one British shorthair, two Devon rex) and one type A random bred cat. CMAH CDS sequence is available and can be found on GeneScaffold ENSFCAG00000002804 ( Primers for the amplification of the coding region of CMAH were designed in both UTRs and intronic regions, flanking the exons. Primers were tested for efficient product amplification on a DNA Engine Gradient Cycler (MJ Research, GMI, Ramsey, MN) and the final PCR conditions were 1.5 mM MgCl2, 1X Denville Buffer, 1 uM primer, 0.2 U Taq and 10 ng DNA. Thermal cycling conditions were as follows: initial denaturation at 94°C for 4 min, followed by 40 cycles of: 94°C for 30 sec, 60°C for 30 sec, 72°C for 45 sec and final extension at 72°C for 20 min. The PCR products were purified with ExoSap (USB, Cleveland, OH) following the manufacturer’s recommendations and directly sequenced using the BigDye terminator Sequencing Kit v3.1 (Applied Biosystems, Foster City, CA). Products were purified with Illustra Sephadex G-50 (Ge Healthcare, Piscataway, NJ) according to manufacturer’s recommendations. Sequence data was analyzed with Sequencer 5.1 (GeneCodes, Ann Arbor, MI, USA).

Variant analysis of feline CMAH

Seven CMAH variants were previously described as concordant with blood type B [4]. Of these, two upstream SNPs (c.1-371C>T and c.1-217G>A), two missense mutations (c.268T>A and c.1603G>A) and one silent variant (c.1269G>A) have been shown not to be associated with blood type B (data not shown) or with a low impact on the protein and not prioritized in the genetic testing. The remaining two variants, a 18 bp deletion (c.1-53delGTCGAAGCCAACGAGCAA) in the 5’UTR and the c.142G>A are predicted to have an impact on the cytidine monophospho-N-acetylneuraminic acid hydroxylase activity and are in linkage disequilibrium with blood type B, and were hence genotyped when possible in all samples included in the study. The c.142G>A variant associated with blood type B [4] was genotyped by direct sequencing using the primer set CMAH-2F and CMAH-2R (S5 Table). The c.142G>A variant is included in the 62,897 SNPs tested on the Illumina array, hence data for the locus was obtained for all GWAS samples by exporting the genotyping results. The 18 bp deletion was genotyped as a size variant by direct PCR using the FAM-CMAHdelF labelled primer and the CMAHdelR primer and electrophoretically separated on an ABI DNA analyzer. The predicted size of the wild-type allele was 319 bp, and 301 bp for the B allele and verified using the program STRand [56]. The PCR conditions were as described: 2 min denaturation at 95°C, followed by 40 cycles at 95°C for 30 sec, 58°C for 30 sec and 72°C for 45 sec, followed by a 10 min extension at 72°C. The newly identified c.364C>T variant was genotyped by direct sequencing using the genomic primers CMAH4F and CMAH4R, and sequenced as previously described in the genomic analysis section (S5 Table).

At the VGL, an allele specific assay was developed for the detection of c.364C>T. Each PCR contained 3μl of template DNA isolated as previously described [57]. Thermal cycling consisted of: 95°C 3 min, 85°C 5 min, 33 cycles of 95°C 45 sec, 60°C 30 sec, 72°C 30 sec, followed by 1 cycle of 72°C for 15 min and a 10°C hold. Each 25 μl reaction contained 1μM of each primer (the forward FAM-364, the wild-type reverse 364wtR and the affected reverse 364affectedR), 1X reaction buffer (Denville Scientific, Inc., South Plainfield, NJ), 2.5mM MgCl2, 1.0mM dNTP, 0.02μl DMSO (Fisher Scientific) and 1.0U Choice Taq DNA polymerase (Denville Scientific Inc.). Template specificity was increased by incorporating a sequence mismatch in the reverse primer (lower-case letter) and the addition of 7 mismatched bases (underlined) created a visually detectable product size variant. To decrease background, template DNA and primers were initially denatured for 3 min at 95°C and the remaining 15 μl of reaction mix was added during the 85°C pre-cycling soak. PCR amplicons were visualized on an ABI3730 DNA analyzer (Applied Biosystems) and analyzed using STRand software [56].

At LVS, pyrosequencing was used to detect c.142G>A, c.139C>T and c.364C>T. Each PCR contained GoTaq Master Mix (Promega, UK), 0.2μM each primer (exon 2 primers for c.142G>A, c.139C>T and exon 4 primers for c.364C>T) (S5 Table) and 2μl of template DNA. Thermal cycling consisted of 95°C for 2 min followed by 36 cycles of 95°C for 20 sec and 58°C for 40 sec, with a final 10°C hold. Amplicons were purified and subjected to pyrosequencing using specific sequencing primers (S5 Table) and PyroGold reagents as described by the manufacturer (Qiagen, UK) on a PyroMark Q24 pyrosequencer (Qiagen).

CMAH computer modelling

A 3-dimensional computer model for the wild-type cat CMAH protein (EF127684.1) was built following a template based modeling procedure [58]. Specifically, PSI-BLAST [30] was used to search the sequence of CMAH against the sequences in the Protein Data Bank (PDB) [59] to find homologous proteins with known tertiary structures. The homologous protein matches were ranked by the significance of sequence similarity (e-value) and the percentage of sequence identity and the top three templates analyzed. Finally, using the 3-dimensional structure of the top ranked protein as a template in conjunction with the sequence alignment, a 3-dimensional structure for the wild-type CMAH protein was built, for the regions for which templates were found, using Modeller [60]. The structural model of the wild-type protein was then used as a template to construct the structural model of the mutated protein.

Supporting Information

S1 Fig. Chromatograms showing the variants in CMAH.

a. The c.142G>A variant associated with blood type B (red rectangle), all possible genotypes are shown. Three bases upstream, in the black rectangle, is shown the c.139C>T variant. b. The c.364C>T variant associated with blood type AB (red rectangle). All possible genotypes are shown.


S2 Fig. Three-dimensional model and templates found by PSI-BLAST for the feline sequence against the PDB database.

(top) The two domains found in the database used to partially modulate the structure of the protein from position 21 to position 94 and from position 192 to position 288 of the CMAH protein. The relative position of the V48M variant associated with blood type B is shown in red in the 2QPZ template and the four zinc-binding sites are shown in orange in the 3JQ4 template. (bottom) Several views of the amino acid substitutions are shown. (a) The residue highlighted in green shows the wild-type amino acid while the residue in light blue show the corresponding mutated residue at position 48 of the protein. (b,c) Different views of the amino acid side chains surrounding the V48M missense substitution.


S1 Table. Details of five highest associated locations of SNPs to CMAH that remained genome-wide significant after permutation testing.


S2 Table. Details of the 5’ UTR deletion, c.139C>T (LVS only samples), c.142G>A and c.364C>T results in all samples with known serology.


S3 Table. DNA SNP analyses of CMAH in type A and AB cats.


S4 Table. Details of VGL genotyping results for all the samples with unknown serology.


S5 Table. Primers sequences for CMAH analysis and genotyping.


Author Contributions

Conceived and designed the experiments: BG LAL. Performed the experiments: BG RAG NG GA CH. Analyzed the data: BG. Contributed reagents/materials/analysis tools: BG DP ES LAL GA CH. Wrote the paper: BG CH. Conducted the 3D protein analysis: BA JC.


  1. 1. Andrews GA, Chavey PS, Smith JE, Rich L (1992) N-glycolylneuraminic acid and N-acetylneuraminic acid define feline blood group A and B antigens. Blood 79: 2485–2491. pmid:1571562
  2. 2. Jourdian GW, Roseman S (1962) The sialic acids. II. Preparation of N-glycolylhexosamines, N-glycolylhexosamine 6-phosphates, glycolyl coenzyme A, and glycolyl glutathione. J Biol Chem 237: 2442–2446. pmid:14452591
  3. 3. Schoop HJ, Schauer R, Faillard H (1969) [On the biosynthesis of N-glycolyneuraminic acid. Oxidative formation of N-glycolylneuraminic acid from N-acetylneuraminic acid]. Hoppe Seylers Z Physiol Chem 350: 155–162. pmid:5776501
  4. 4. Bighignoli B, Niini T, Grahn RA, Pedersen NC, Millon LV, Polli M, et al. (2007) Cytidine monophospho-N-acetylneuraminic acid hydroxylase (CMAH) mutations associated with the domestic cat AB blood group. BMC Genet 8: 27. pmid:17553163
  5. 5. Hayakawa T, Aki I, Varki A, Satta Y, Takahata N (2006) Fixation of the human-specific CMP-N-acetylneuraminic acid hydroxylase pseudogene and implications of haplotype diversity for human evolution. Genetics 172: 1139–1146. pmid:16272417
  6. 6. Chou HH, Takematsu H, Diaz S, Iber J, Nickerson E, Wright KL, et al. (1998) A mutation in human CMP-sialic acid hydroxylase occurred after the Homo-Pan divergence. Proc Natl Acad Sci U S A 95: 11751–11756. pmid:9751737
  7. 7. Springer SA, Diaz SL, Gagneux P (2014) Parallel evolution of a self-signal: humans and new world monkeys independently lost the cell surface sugar Neu5Gc. Immunogenetics 66: 671–674. pmid:25124893
  8. 8. Schauer R, Srinivasan GV, Coddeville B, Zanetta JP, Guerardel Y (2009) Low incidence of N-glycolylneuraminic acid in birds and reptiles and its absence in the platypus. Carbohydr Res 344: 1494–1500. pmid:19541293
  9. 9. Segurel L, Thompson EE, Flutre T, Lovstad J, Venkat A, Margulis SW, et al. (2012) The ABO blood group is a trans-species polymorphism in primates. Proc Natl Acad Sci U S A 109: 18493–18498. pmid:23091028
  10. 10. Varki A (2001) Loss of N-glycolylneuraminic acid in humans: Mechanisms, consequences, and implications for hominid evolution. Am J Phys Anthropol Suppl 33: 54–69.
  11. 11. Martin MJ, Rayner JC, Gagneux P, Barnwell JW, Varki A (2005) Evolution of human-chimpanzee differences in malaria susceptibility: relationship to human genetic loss of N-glycolylneuraminic acid. Proc Natl Acad Sci U S A 102: 12819–12824. pmid:16126901
  12. 12. Rich SM, Leendertz FH, Xu G, LeBreton M, Djoko CF, Aminake MN, et al. (2009) The origin of malignant malaria. Proc Natl Acad Sci U S A 106: 14902–14907. pmid:19666593
  13. 13. Auer L, Bell K (1981) The AB blood group system of cats. Anim Blood Groups Biochem Genet 12: 287–297. pmid:7342803
  14. 14. Griot-Wenk ME, Callan MB, Casal ML, Chisholm-Chait A, Spitalnik SL, Patterson DF, et al. (1996) Blood type AB in the feline AB blood group system. Am J Vet Res 57: 1438–1442. pmid:8896680
  15. 15. Giger U, Bucheler J, Patterson DF (1991) Frequency and inheritance of A and B blood types in feline breeds of the United States. J Hered 82: 15–20. pmid:1997588
  16. 16. Griot-Wenk ME, Giger U (1995) Feline transfusion medicine. Blood types and their clinical importance. Vet Clin North Am Small Anim Pract 25: 1305–1322. pmid:8619268
  17. 17. Forcada Y, Guitian J, Gibson G (2007) Frequencies of feline blood types at a referral hospital in the south east of England. J Small Anim Pract 48: 570–573. pmid:17608664
  18. 18. Giger U, Kilrain CG, Filippich LJ, Bell K (1989) Frequencies of feline blood groups in the United States. J Am Vet Med Assoc 195: 1230–1232. pmid:2584120
  19. 19. Arikan S, Gurkan M, Ozaytekin E, Dodurka T, Giger U (2006) Frequencies of blood type A, B and AB in non-pedigree domestic cats in Turkey. J Small Anim Pract 47: 10–13. pmid:16417604
  20. 20. Tasker S, Barker EN, Day MJ, Helps CR (2014) Feline blood genotyping versus phenotyping, and detection of non-AB blood type incompatibilities in UK cats. J Small Anim Pract 55: 185–189. pmid:24697343
  21. 21. Proverbio D, Spada E, Baggiani L, Perego R, Milici A, Ferro E (2011) Comparison of gel column agglutination with monoclonal antibodies and card agglutination methods for assessing the feline AB group system and a frequency study of feline blood types in northern Italy. Vet Clin Pathol 40: 32–39. pmid:21299585
  22. 22. Proverbio D, Spada E, Perego R, Della Pepa A, Bagnagatti De Giorgi G, Baggiani L (2013) Assessment of blood types of Ragdoll cats for transfusion purposes. Vet Clin Pathol 42: 157–162. pmid:23654225
  23. 23. Ejima H, Kurokawa K, Ikemoto S (1986) Feline red blood cell groups detected by naturally occurring isoantibody. Nihon Juigaku Zasshi 48: 971–976. pmid:3784230
  24. 24. Silvestre-Ferreira AC, Pastor J, Almeida O, Montoya A (2004) Frequencies of feline blood types in northern Portugal. Vet Clin Pathol 33: 240–243. pmid:15570562
  25. 25. Medeiros MA, Soares AM, Alviano DS, Ejzemberg R, da Silva MH, Almosny NR (2008) Frequencies of feline blood types in the Rio de Janeiro area of Brazil. Vet Clin Pathol 37: 272–276. pmid:18761518
  26. 26. Malik R, Griffin DL, White JD, Rozmanec M, Tisdall PL, Foster SF, et al. (2005) The prevalence of feline A/B blood types in the Sydney region. Aust Vet J 83: 38–44. pmid:15971816
  27. 27. Fosset FT, Blais MC (2014) Prevalence of feline blood groups in the Montreal area of Quebec, Canada. Can Vet J 55: 1225–1228. pmid:24381340
  28. 28. Bucheler J, Giger U (1993) Alloantibodies against A and B blood types in cats. Vet Immunol Immunopathol 38: 283–295. pmid:8291206
  29. 29. Giger U, Bucheler J (1991) Transfusion of type-A and type-B blood to cats. J Am Vet Med Assoc 198: 411–418. pmid:2010334
  30. 30. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, et al. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402. pmid:9254694
  31. 31. Johnson DC, Dean DR, Smith AD, Johnson MK (2005) Structure, function, and formation of biological iron-sulfur clusters. Annu Rev Biochem 74: 247–281. pmid:15952888
  32. 32. Montague MJ, Li G, Gandolfi B, Khan R, Aken BL, Searle SM, et al. (2014) Comparative analysis of the domestic cat genome reveals genetic signatures underlying feline biology and domestication. Proc Natl Acad Sci U S A 111: 17230–17235. pmid:25385592
  33. 33. Gandolfi B, Gruffydd-Jones TJ, Malik R, Cortes A, Jones BR, Helps CR, et al. (2012) First WNK4-hypokalemia animal model identified by genome-wide association in Burmese cats. PLoS One 7: e53173. pmid:23285264
  34. 34. Alhaddad H, Gandolfi B, Grahn RA, Rah HC, Peterson CB, Maggs DJ, et al. (2014) Genome-wide association and linkage analyses localize a progressive retinal atrophy locus in Persian cats. Mamm Genome.
  35. 35. Lyons LA, Erdman CA, Grahn RA, Hamilton MJ, Carter MJ, Helps CR, et al. (2016) Aristaless-Like Homeobox protein 1 (ALX1) variant associated with craniofacial structure and frontonasal dysplasia in Burmese cats. Developmental Biology 409: 451–458. pmid:26610632
  36. 36. Gandolfi B, Alhaddad H, Joslin SE, Khan R, Filler S, Brem G, et al. (2013) A splice variant in KRT71 is associated with curly coat phenotype of Selkirk Rex cats. Sci Rep 3: 2000. pmid:23770706
  37. 37. Gandolfi B, Alhaddad H, Affolter VK, Brockman J, Haggstrom J, Joslin SE, et al. (2013) To the Root of the Curl: A Signature of a Recent Selective Sweep Identifies a Mutation That Defines the Cornish Rex Cat Breed. PLoS One 8: e67105. pmid:23826204
  38. 38. Lipinski MJ, Froenicke L, Baysac KC, Billings NC, Leutenegger CM, Levy AM, et al. (2008) The ascent of cat breeds: genetic evaluations of breeds and worldwide random-bred populations. Genomics 91: 12–21. pmid:18060738
  39. 39. Kurushima JD, Lipinski MJ, Gandolfi B, Froenicke L, Grahn JC, Grahn RA, et al. (2013) Variation of cats under domestication: genetic assignment of domestic cats to breeds and worldwide random-bred populations. Anim Genet 44: 311–324. pmid:23171373
  40. 40. Drogemuller C, Rufenacht S, Wichert B, Leeb T (2007) Mutations within the FGF5 gene are associated with hair length in cats. Anim Genet 38: 218–221. pmid:17433015
  41. 41. Lyons LA, Imes DL, Rah HC, Grahn RA (2005) Tyrosinase mutations associated with Siamese and Burmese patterns in the domestic cat (Felis catus). Anim Genet 36: 119–126. pmid:15771720
  42. 42. Alhaddad H, Khan R, Grahn RA, Gandolfi B, Mullikin JC, Cole SA, et al. (2013) Extent of linkage disequilibrium in the domestic cat, Felis silvestris catus, and its breeds. PLoS One 8: e53537. pmid:23308248
  43. 43. Yang C, Pring M, Wear MA, Huang M, Cooper JA, Svitkina TM, et al. (2005) Mammalian CARMIL inhibits actin filament capping by capping protein. Dev Cell 9: 209–221. pmid:16054028
  44. 44. Tangvoranuntakul P, Gagneux P, Diaz S, Bardor M, Varki N, Varki A, et al. (2003) Human uptake and incorporation of an immunogenic nonhuman dietary sialic acid. Proc Natl Acad Sci U S A 100: 12045–12050. pmid:14523234
  45. 45. Varki A (2010) Colloquium paper: uniquely human evolution of sialic acid genetics and biology. Proc Natl Acad Sci U S A 107 Suppl 2: 8939–8946. pmid:20445087
  46. 46. Julien S, Adriaenssens E, Ottenberg K, Furlan A, Courtand G, Vercoutter-Edouart AS, et al. (2006) ST6GalNAc I expression in MDA-MB-231 breast cancer cells greatly modifies their O-glycosylation pattern and enhances their tumourigenicity. Glycobiology 16: 54–64. pmid:16135558
  47. 47. Mortezai N, Behnken HN, Kurze AK, Ludewig P, Buck F, Meyer B, et al. (2013) Tumor-associated Neu5Ac-Tn and Neu5Gc-Tn antigens bind to C-type lectin CLEC10A (CD301, MGL). Glycobiology 23: 844–852. pmid:23507963
  48. 48. Kohn B, Niggemeier A, Reitemeyer S, Giger U (1997) Blood typing in cats with a new test card method. Kleintierpraxis 42: 941-+.
  49. 49. Lapierre Y, Rigal D, Adam J, Josef D, Meyer F, Greber S, et al. (1990) The Gel Test—a New Way to Detect Red-Cell Antigen-Antibody Reactions. Transfusion 30: 109–113. pmid:2305438
  50. 50. Stieger K, Palos H, Giger U (2005) Comparison of various blood-typing methods for the feline AB blood group system. American Journal of Veterinary Research 66: 1393–1399. pmid:16173483
  51. 51. Seth M, Jackson KV, Giger U (2011) Comparison of five blood-typing methods for the feline AB blood group system. Am J Vet Res 72: 203–209. pmid:21281194
  52. 52. Spada E, Proverbio D, Baggiani L, Bagnagatti De Giorgi G, Perego R, Ferro E (2015) Evaluation of an immunochromatographic test for feline AB system blood typing. J Vet Emerg Crit Care (San Antonio).
  53. 53. Chenu S, Gregoire A, Malykh Y, Visvikis A, Monaco L, Shaw L, et al. (2003) Reduction of CMP-N-acetylneuraminic acid hydroxylase activity in engineered Chinese hamster ovary cells using an antisense-RNA strategy. Biochim Biophys Acta 1622: 133–144. pmid:12880951
  54. 54. Malykh YN, Shaw L, Schauer R (1998) The role of CMP-N-acetylneuraminic acid hydroxylase in determining the level of N-glycolylneuraminic acid in porcine tissues. Glycoconj J 15: 885–893. pmid:10052592
  55. 55. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. (2007) PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 81: 559–575. pmid:17701901
  56. 56. Toonen RJ, Hughes S (2001) Increased throughput for fragment analysis on an ABI PRISM 377 automated sequencer using a membrane comb and STRand software. Biotechniques 31: 1320–1324. pmid:11768661
  57. 57. Oberbauer AM, Grossman DI, Eggleston ML, Irion DN, Schaffer AL, Pedersen NC, et al. (2003) Alternatives to blood as a source of DNA for large-scale scanning studies of canine genome linkages. Vet Res Commun 27: 27–38. pmid:12625401
  58. 58. Cheng J, Li J, Wang Z, Eickholt J, Deng X (2012) The MULTICOM toolbox for protein structure prediction. BMC Bioinformatics 13: 65. pmid:22545707
  59. 59. Berman H, Henrick K, Nakamura H (2003) Announcing the worldwide Protein Data Bank. Nat Struct Biol 10: 980. pmid:14634627
  60. 60. Eswar N, Eramian D, Webb B, Shen MY, Sali A (2008) Protein structure modeling with MODELLER. Methods Mol Biol 426: 145–159. pmid:18542861