Genetic variation in the major histocompatibility complex of the European brown hare (Lepus europaeus) across distinct phylogeographic areas

The major histocompatibility complex is one of the best studied systems in vertebrates providing evidence for the long-term action of selection. Here, we examined the intra- and inter-population genetic diversity of the MHC class II DRB locus in European brown hare (Lepus europaeus) and correlated the results with genetic variability already estimated from the MHC DQA locus and from maternally (mitochondrial DNA (mtDNA)) and biparentally (allozymes, microsatellites) inherited loci. L. europaeus showed remarkable genetic polymorphism in both DQA and DRB1 loci. The Anatolian populations exhibited the highest genetic polymorphism for both loci. Balancing selection has established increased variability in the European populations despite the founder effects after the last glaciation. Different evolutionary rates were traced for DRB1 and DQA loci, as evidenced by the higher number of common DRB1 than DQA alleles and the greater differences between DRB1 alleles with common origin in comparison with DQA alleles. The high number of rare alleles with low frequencies detected implies that frequency-dependent selection drives MHC evolution in the brown hare through the advantage of rare alleles. Both loci were under the influence of positive selection within the peptide-binding region. The functional polymorphism, recorded as amino acid substitutions within the binding pockets, fell also within distinct geographic patterns, yet it was much narrower than the genetic polymorphism. We hypothesize that certain structural and functional characteristics of the binding pockets set limitations to the actual shape of genetic polymorphism in MHC.     

Analysis of genetic diversity at the DQA loci in African cattle: evidence for a BoLA-DQA3 locus

 We describe the development of a polymerase chain reaction (PCR)-based approach for analysis of genetic diversity at the DQA loci in African Bos indicus and Bos taurus cattle. This approach, equally effective in European and Asian cattle breeds, detects the presence or absence of DQA1 and most duplicated DQA2 genes. Nucleotide and predicted amino acid sequence analysis of the highly polymorphic second exons, in addition to analysis of the locus-specific and relatively non-polymorphic transmembrane, cytoplasmic, and 3-prime untranslated regions, has provided evidence for considerable diversity between each of the duplicated DQA2 genes. Therefore, we propose the designation BoLA-DQA3 for the previously unpublished alleles at the second DQA2 locus. Fourteen distinct PCR restriction fragment length polymorphism (RFLP) patterns, each identifying families of alleles at three DQA loci, can be distinguished. Nucleotide sequence analysis of new PCR-RFLP patterns from 193 Kenyan Boran, Ethiopian Arsi (B. indicus), and Guinean N’Dama (B. taurus) cattle identified 13 DQA1 alleles within eight major allelic families, five DQA2 alleles within a single allelic family, and seven DQA3 alleles within three major allelic families. Received: 19 February 1997 / Revised: 28 February 1997     

Genomic diversity and selection sweeps identified in Indian swamp buffaloes reveals it's uniqueness with riverine buffaloes

The present investigation was focused to study genomic diversity of Indian swamp buffalo populations through reduced representation approach (ddRAD). The heterozygosity (F_ST) among the swamp buffaloes was 0.11 between Assam and Manipuri; 0.20 between swamp (Manipuri) and riverine buffaloes; 0.30 between swamp (Manipuri) and cattle. The average observed and expected heterozygosity in swamp buffalo populations was 0.254 and 0.221 respectively. The Inbreeding coefficient (F_IS) value was 0.02 among the swamp buffaloes. PCA and structure analysis revealed Manipuri swamp buffalo was genetically distinct and closely related to Nagaland swamp buffalo and least to Assam swamp buffalo. Identification of selective sweeps revealed 1087 regions to have undergone selection related to immune response, adaptation and nervous system. A total of 3451 SSRs were identified in the genome of swamp buffaloes. The study evidenced the genomic diversity in the swamp buffalo populations and its uniqueness in comparison with riverine buffalo and cattle.     

The transcription of MGAT4A glycosyl transferase is increased in white cells of peripheral blood of Type 2 Diabetes patients

Background

The giant panda (Ailuropoda melanoleuca) is one of the most endangered animals due to habitat fragmentation and loss. Although the captive breeding program for this species is now nearly two decades old, researches on the genetic background of such captive populations, especially on adaptive molecular polymorphism of major histocompatibility complex (MHC), are still limited. In this study, we characterized adaptive variation of the giant panda's MHC DQA gene by PCR amplification of its antigen-recognizing region (i.e. the exon 2) and subsequent single-strand conformational polymorphism (SSCP) and sequence analyses.

Results

The results revealed a low level of DQA exon 2 diversity in this rare animal, presenting 6 alleles from 61 giant panda individuals. The observed polymorphism was restricted to 9 amino acid substitutions, all of which occurred at and adjacent to positions forming the functionally important antigen-binding sites. All the samples were in Hardy-Weinberg proportions. A significantly higher rate of non-synonymous than synonymous substitutions at the antigen-binding sites indicated positive selection for diversity in the locus.

Conclusion

The DQA allelic diversity of giant pandas was low relative to other vertebrates. Nonetheless, the pandas exhibited more alleles in DQA than those in DRB, suggesting the alpha chain genes would play a leading role when coping with certain pathogens and thus should be included in conservation genetic investigation. The microsatellite and MHC loci might predict long-term persistence potential and short-term survival ability, respectively. Consequently, it is recommended to utilize multiple suites of microsatellite markers and multiple MHC loci to detect overall genetic variation in order to design unbiased conservation strategies.     

Identification of Mamu-DPA1, Mamu-DQA1, and Mamu-DRA alleles in a cohort of Chinese rhesus macaques

Rhesus macaques have long been used as animal models for various human diseases; the susceptibility and/or resistance to some of these diseases are related to the major histocompatibility complex (MHC). To gain insight into the MHC background and to facilitate the experimental use of Chinese rhesus macaques, Mamu-DPA1, Mamu-DQA1, and Mamu-DRA alleles were investigated in 30 Chinese rhesus macaques by gene cloning and sequencing. A total of 14 Mamu-DPA1, 17 Mamu-DQA1, and 9 Mamu-DRA alleles were identified in this study. Of these alleles, 22 novel sequences have not been documented in earlier studies, including nine Mamu-DPA1, ten Mamu-DQA1, and three Mamu-DRA alleles. Interestingly, like Mafa-DQA1 and Mafa-DPA1, more than two Mamu-DQA1 and Mamu-DPA1 alleles were detected in one animal in this study, which suggested that they might represent gene duplication. If our findings can be validated by other studies, it will further increase the number of known Mamu-DPA1 and Mamu-DQA1 polymorphisms. Our data also indicated significant differences in MHC class II allele distribution among the Chinese rhesus macaques, Vietnamese cynomolgus macaques, and the previously reported rhesus macaques, which were mostly of Indian origin. This information will not only promote the understanding of Chinese rhesus macaque MHC diversity and polymorphism but will also facilitate the use of Chinese rhesus macaques in studies of human disease.     

Major histocompatibility complex class II genetic variation in forest musk deer (Moschus berezovskii) in China

The major histocompatibility complex (MHC) plays an important role in the immune system of vertebrates. We used the second exon of four MHC class II genes (DRA, DQA1, DQA2 and DRB3) to assess the overall MHC variation in forest musk deer (Moschus berezovskii). We also compared the MHC variation in captive and wild populations. We observed 22 alleles at four loci (four at DRA, four at DQA1, four at DQA2 and 10 at DRB3), 15 of which were newly identified alleles. Results suggest that forest musk deer maintain relatively high MHC variation, which may result from balancing selection. Moreover, considerable diversity was observed at the DRA locus. We found a high frequency of Mobe‐DRA*02, Mobe‐DQA1*01 and Mobe‐DQA2*05 alleles, which may be important for pathogen resistance. A Ewens–Watterson test showed that the DRB3 locus in the wild population had experienced recent balancing selection. We detected a small divergence at the DRA locus, suggesting the effect of weak positive selection on the DRA gene. Alternatively, this locus may be young and not yet adapted a wide spectrum of alleles for pathogen resistance. The significant heterozygosity deficit observed at the DQA1 and DRB3 loci in the captive population and at all four loci in the wild population may be the result of a population bottleneck. Additionally, MHC genetic diversity was higher in the wild population than in the captive, suggesting that the wild population may have the ability to respond to a wider range of pathogens.     

Mitochondrial sequence-based evolutionary analysis of riverine–swamp hybrid buffaloes of India indicates novel maternal differentiation and domestication patterns

In this study, mitochondrial D-loop sequence data on riverine, swamp and hybrid buffaloes from India have been generated and compared with other reported Indian riverine, Chinese and Bangladeshi swamp buffalo populations. Sequence analysis revealed the presence of 132 haplotypes, with a haplotype diversity of 0.9611 ± 0.0045 and a nucleotide diversity of 0.04801 ± 0.00126. For the first time, the existence of riverine–swamp hybrids among the Indian Chilika buffalo population has been recorded, having 49 chromosomes, which was also confirmed by mitochondrial haplotype sharing between Chilika and Indian swamp as well as Chinese swamp buffalo populations in the network analysis. Phylogenetic analysis documents the sharing of reported pre-domestication haplogroups ‘SA1’, ‘SA2’, ‘SA3’ and ‘SB1’ between the Chilika and swamp buffalo populations of India, China and Bangladesh, an indication of the migration of swamp buffaloes towards Bangladesh and adjoining lower parts of India and north towards Chinese domestication sites. The results have also been supplemented by multidimension scaling, grouping Indian and Chinese swamp buffaloes more closely together with Bangladeshi buffaloes, but into a separate quadrant, whereas Chilika grouped away from other riverine as well as swamp buffaloes. These findings thus confirm the previous reports that the northeast region of India, close to the Indo-China border, is the point of evolution of swamp buffaloes with multiple sites of domestication.     

Allelic diversity at the Mhc-DQA locus of woodmouse populations (Apodemus sylvaticus) present in the islands and mainland of the northern Mediterranean

Aim Polymorphism at neutral markers and at MHC loci in rodent populations living on islands is generally low. The main genetic factors that may contribute to a reduced level of genetic variability are genetic drift, reduced gene flow and founder events. We investigated the pattern of polymorphism at the second exon of the Mhc‐DQA gene in island populations of Apodemus sylvaticus and in their mainland counterparts to investigate the pattern of MHC polymorphism in a phylogeographical context and to assess the impact of insularity on diversity at this locus. Location Eight north Mediterranean populations of Apodemus sylvaticus were studied, including five island populations (Majorca, Minorca, Porquerolles, Port‐Cros and Sicily) and three mainland populations. Methods cDNA sequencing and nucleotide sequences analyses. Synonymous and non‐synonymous substitutions were examined at the PBR and non‐PBR sites. The DQA allelic distribution in populations was compared with the woodmouse phylogeography. Results This study presents novel DQA alleles. High polymorphism of the DQA locus is recorded in natural populations of A. sylvaticus with 13 alleles being widely distributed irrespective of the geographical origin and palaeoclimatic history of populations. The DQA locus does not show the expected pattern for non‐synonymous substitutions at the PBR sites. However, island populations show a weak loss of polymorphism in comparison with their mainland counterparts. Main conclusions The DQA locus in the woodmouse seems to be subject to weak selection and does not allow resolution of phylogeographical relationships among European woodmouse populations. The presence of at least three alleles in island populations and the maintenance of five alleles between the two European lineages over 1.5 Myr suggest that balancing selection may act within populations, and more precisely within island populations, to maintain genetic variability. This study shows that phylogeographical studies are a prerequisite for any genetic investigation of selected genes in natural populations.     

Mapping and characterization of the DQ subregion of the ovine MHC

A map of the ovine MHC class II DQ subregion has been constructed from overlapping cosmid clones. This region consists of two loci linked on a linear tract of 130 kb DNA. Each locus consists of a DQA and a DQB gene in a tail-to-tail orientation. The genes in each locus are transcribed but only those designated DQ1 express class II molecules at the surface of mouse L cells following DNA-mediated gene transfection. The DQA1 and DQB1 genes are separated by 11kb while the DQA2 and B2 genes are 25 kb apart. The loci are separated by 22 kb.     

Allelic diversity at the Mhc-DQA locus in cotton rats (Sigmodon hispidus) and a comparison of DQA sequences within the family Muridae (Mammalia: Rodentia)

 The cotton rat (Sigmodon hispidus) is a common murid rodent of the southern United States, Mexico, and Central America. Using single-stranded conformation polymorphism analysis and DNA sequencing techniques, 11 DQA exon 2 alleles were detected among 180 S. hispidus from Caddo County, Oklahoma, USA. The alleles represent a single locus exhibiting a high level of polymorphism. Nucleotide and amino acid distance values among DQA alleles of S. hispidus were higher than those within Mus musculus and species of Rattus. Although the distribution of polymorphic amino acid residues among alleles of S. hispidus was similiar to that of Mus and Rattus, some residues of the α-helix region were more variable in S. hispidus. Comparisons of nonsynonymous and synonymous substitutions indicated a trend toward higher numbers of nonsynonymous substitutions; however, this difference was not significant statistically among S. hispidus alleles. To examine evolution of DQA alleleswithin Muridae, we performed a phylogenetic analysis that included DQA alleles from S. hispidus, Peromyscus leucopus, M. musculus, R. norvegicus, and six Australian species of Rattus. Results depicted monophyly for each genus, and this concordance between species and gene trees represents a lack of evidence for trans-species persistence of alleles among these genera. Received: 22 October 1998 / Revised: 17 March 1999     

Polymorphisms in MHC-DRA and -DRB alleles of water buffalo (Bubalus bubalis) reveal different features from cattle DR alleles

Seventy-five individuals of Bubalus bubalis belonging to four different breeds, three of river buffalo and one of swamp buffalo, were studied for polymorphism in MHC DRB (Bubu-DRB) and DRA (Bubu-DRA) loci. Eight alleles of Bubu-DRB were found, and all alleles in the swamp type were shared with the three river breeds. All alleles sampled from the breed of European origin (Mediterranean) were present in breeds sampled in Brazil, thus variability of this locus may have been preserved to a great extent in the more recently founded Brazilian population. Bubu-DRB alleles contained higher proportions of synonymous vs. non-synonymous substitutions in the non-peptide-binding sites (PBS) region, in contrast to the pattern of variation found in BoLA-DRB3, the orthologous locus in cattle. This indicated that either the first domain exon (exon 2) of Bubu-DRB has not undergone as much recombination and/or gene conversion as in cattle alleles, or Bubu-DRB may be more ancient than BoLA-DRB3 alleles. Phylogenetic analysis of DRB alleles from Bubalus, Syncerus c. caffer, the Cape buffalo, and domestic cattle demonstrated transspecies polymorphism. Water buffalo contained two alleles of DRA that differed from each other in two amino acid positions, including one in the PBS (alpha22) that was also shared with Anoa depressicornis, the anoa. Discovery of variation in DRA was surprising as the first domain of DRA is a highly conserved polypeptide in mammals in general and especially in ruminants, where no other substitution in PBS was seen.     

Major histocompatibility complex variation at three class II loci in the northern elephant seal

Northern elephant seals were hunted to near extinction in the 19th century, yet have recovered remarkably and now number around 175,000. We surveyed 110 seals for single-strand conformation polymorphism (SSCP) and sequence variation at three major histocompatibility (MHC) class II loci (DQA, DQB and DRB) to evaluate the genetic consequences of the population bottleneck at these loci vs. other well-studied genes. We found very few alleles at each MHC locus, significant variation among breeding sites for the DQA locus, and linkage disequilibrium between the DQB and DRB loci. Northern elephant seals are evidently inbred, although there is as yet no evidence of correlative reductions in fitness.     

Isolation, characterization and evolution of ovine major histocompatibility complex class II DRA and DQA genes

Four full-length ovine major histocompatibility complex (MHC) class II A cDNA clones coding for new alleles of DRA, DQA1 and DQA2 genes were isolated from two ovine Λgt10 cDNA libraries. The derived amino acid sequences of these clones resemble class II A molecules from other species in both size and structure. Restriction fragment length polymorphism analysis, using an Ovar-DRA probe on DNA from Merino and Romney sheep revealed only limited polymorphism in contrast to the high levels of polymorphism revealed by Ovar-DQA probes. Comparison of the predicted amino acid sequences for the three ovine A genes with class II A genes from five other species revealed that the most variable region of the molecule is the signal peptide. Although virtually every amino acid site shows variation, within or between species, there are some blocks of highly conserved residues. Within gene comparisons of nucleotide differences reveal that the greatest number of changes is found between the alleles of Ovar-DQA1 and -DQA2 genes and the least between Ovar-DRA1 alleles. Phylogenetic analysis of class IIA sequences from several species place DRA and DQA genes on two distinct branches, with Ovar-DRA1 and BOLA-DRA, and Ovar-DQA1 and BOLA-DQA being most similar on their respective branches.     

Comprehensive identification of high-frequency and co-occurring Mafa-DPA1, Mafa-DQA1, Mafa-DRA, and Mafa-DOA alleles in Vietnamese cynomolgus macaques

High-frequency alleles and/or co-occurring human leukocyte antigen alleles across loci appear to be more important than individual alleles as markers of disease risk and have clinical value as biomarkers for targeted screening or the development of new disease therapies. To better elucidate the major histocompatibility complex (MHC) background and to facilitate the experimental use of cynomolgus macaques, Mafa-DPA1, Mafa-DQA1, Mafa-DRA, and Mafa-DOA alleles were characterized, and their combinations were investigated in 30 Vietnamese macaques by gene cloning and sequencing. A total of 26 Mafa-DPA1, 18 Mafa-DQA1, 9 Mafa-DRA, and 15 Mafa-DOA alleles, including 7 high-frequency alleles, were identified in this study, respectively. In addition, 15 Mafa-DQA1, 17 Mafa-DPA1, 15 Mafa-DOA, and 2 Mafa-DRA alleles represented novel sequences that had not been documented in earlier studies. Our results also showed that the Vietnamese macaques might be valuable because no less than 30 % of the test animals possessed Mafa-DRA*01:02:01 (90 %), -DQA1*26:01:03 (37 %), -DOA*01:02:07 (34 %), and -DQA1*01:03:03 (30 %). We previously reported that the combinations of MHC class II alleles, including the combination of DOA*01:02:07-DPA1*02:09 and DOA*01:02:07-DQA1*01:03:03, were detected in 17 and 14 % of the animals, respectively. Interestingly, more than two Mafa-DQA1 and Mafa-DPA1 alleles were detected in one animal in this study, which suggested that they might be caused by a chromosomal duplication. If our findings can be validated by other studies, it will further enrich the number of known Mafa-DPA1 and Mafa-DQA1 polymorphisms. Our results identified the co-occurring MHC alleles across loci in a cohort of Vietnamese cynomolgus macaques, which emphasized the value of this species as a model for biomedical research.     

Polymorphism at the ovine major histocompatibility complex class II loci

Southern hybridization analysis of the ovine major histocompatibility complex (MHC) ( MhcOvar ) class II region, using sheep-specific probes for the DQA1, DQA2, DQB and DRA loci, has revealed extensive polymorphism. DQA1 and DQAP had eight and 16 alleles respectively, DQB had six and DRA had three alleles. Little information was derived from the DRB locus owing to extensive cross-hybridization between the DRB probe and the DQB locus. Differences in allele frequency between breeds were revealed. At the DQA1 locus a null allele (DQA1-N) was observed with a frequency of between 27% and 45%, making this the most common DQA1 allele in all breeds examined. The frequency of DQA1-N homozygotes was between 11% and 18%, raising questions as to the functional significance of the DQA1 gene. Linkage analysis between the DQA1, DQA2, DQB and DRA loci did not reveal any recombination.     

Major histocompatibility complex variation at class II DQA locus in the brown hare (Lepus europaeus)

The major histocompatability complex (MHC) is a multigene family of receptors that bind and present antigenic peptides to T-cells. Genes of the MHC are characterized by an outstanding genetic polymorphism, which is considered to be maintained by positive selection. Sites involved in peptide binding form binding pockets (P) that are collectively termed the peptide-binding region (PBR). In this study, we examined the level of MHC genetic diversity within and among natural populations of brown hare ( Lepus europaeus ) from Europe and Anatolia choosing for analysis of the second exon of the DQA locus, one of the most polymorphic class II loci. We aimed at an integrated population genetic analysis of L. europeaus by (i) correlating MHC polymorphism to genetic variability and phylogenetic status estimated previously from maternally (mtDNA) and biparentally (allozymes, microsatellites) inherited loci; and (ii) comparing full-length exon amino acid polymorphism with functional polymorphism in the PBR and the binding pockets P1, P6 and P9. A substantial level of DQA exon 2 polymorphism was detected with two completely different set of alleles between the Anatolian and European populations. However, the phylogeny of full-length exon 2 Leeu-DQA alleles did not show a strong phylogeographic signal. The presence of balancing selection was supported by a statistically significant excess of nonsynonymous substitutions over synonymous in the PBR and a trans-species pattern of evolution detected after phylogenetic reconstruction. The differentiating patterns detected between genetic and functional polymorphism, i.e. the number and the distribution of pocket variants within and among populations, indicated a hierarchical action of selection pressures.     

Major histocompatibility complex variation and evolution at a single, expressed DQA locus in two genera of elephants

Genes of the vertebrate major histocompatibility complex (MHC) are crucial to defense against infectious disease, provide an important measure of functional genetic diversity, and have been implicated in mate choice and kin recognition. As a result, MHC loci have been characterized for a number of vertebrate species, especially mammals; however, elephants are a notable exception. Our study is the first to characterize patterns of genetic diversity and natural selection in the elephant MHC. We did so using DNA sequences from a single, expressed DQA locus in elephants. We characterized six alleles in 30 African elephants (Loxodonta africana) and four alleles in three Asian elephants (Elephas maximus). In addition, for two of the African alleles and three of the Asian alleles, we characterized complete coding sequences (exons 1–5) and nearly complete non-coding sequences (introns 2–4) for the class II DQA loci. Compared to DQA in other wild mammals, we found moderate polymorphism and allelic diversity and similar patterns of selection; patterns of non-synonymous and synonymous substitutions were consistent with balancing selection acting on the peptides involved in antigen binding in the second exon. In addition, balancing selection has led to strong trans-species allelism that has maintained multiple allelic lineages across both genera of extant elephants for at least 6 million years. We discuss our results in the context of MHC diversity in other mammals and patterns of evolution in elephants.     

Domestication does not narrow MHC diversity in Sus scrofa

The Major Histocompatibility Complex (MHC) is a multigene family of outstanding polymorphism. MHC molecules bind antigenic peptides in the peptide-binding region (PBR) that consists of five binding pockets (P). In this study, we compared the genetic diversity of domestic pigs to that of the modern representatives of their wild ancestors, the wild boar, in two MHC loci, the oligomorphic DQA and the polymorphic DRB1. MHC nucleotide polymorphism was compared with the actual functional polymorphism in the PBR and the binding pockets P1, P4, P6, P7, and P9. The analysis of approximately 200 wild boars collected throughout Europe and 120 domestic pigs from four breeds (three pureblood, Pietrain, Leicoma, and Landrace, and one mixed Danbred) revealed that wild boars and domestic pigs share the same levels of nucleotide and amino acid polymorphism, allelic richness, and heterozygosity. Domestication did not appear to act as a bottleneck that would narrow MHC diversity. Although the pattern of polymorphism was uniform between the two loci, the magnitude of polymorphism was different. For both loci, most of the polymorphism was located in the PBR region and the presence of positive selection was supported by a statistically significant excess of nonsynonymous substitutions over synonymous substitutions in the PBR. P4 and P6 were the most polymorphic binding pockets. Functional polymorphism, i.e., the number and the distribution of pocket variants within and among populations, was significantly narrower than genetic polymorphism, indicative of a hierarchical action of selection pressures on MHC loci.     

Imputing Amino Acid Polymorphisms in Human Leukocyte Antigens

DNA sequence variation within human leukocyte antigen (HLA) genes mediate susceptibility to a wide range of human diseases. The complex genetic structure of the major histocompatibility complex (MHC) makes it difficult, however, to collect genotyping data in large cohorts. Long-range linkage disequilibrium between HLA loci and SNP markers across the major histocompatibility complex (MHC) region offers an alternative approach through imputation to interrogate HLA variation in existing GWAS data sets. Here we describe a computational strategy, SNP2HLA, to impute classical alleles and amino acid polymorphisms at class I (HLA-A, -B, -C) and class II (-DPA1, -DPB1, -DQA1, -DQB1, and -DRB1) loci. To characterize performance of SNP2HLA, we constructed two European ancestry reference panels, one based on data collected in HapMap-CEPH pedigrees (90 individuals) and another based on data collected by the Type 1 Diabetes Genetics Consortium (T1DGC, 5,225 individuals). We imputed HLA alleles in an independent data set from the British 1958 Birth Cohort (N = 918) with gold standard four-digit HLA types and SNPs genotyped using the Affymetrix GeneChip 500 K and Illumina Immunochip microarrays. We demonstrate that the sample size of the reference panel, rather than SNP density of the genotyping platform, is critical to achieve high imputation accuracy. Using the larger T1DGC reference panel, the average accuracy at four-digit resolution is 94.7% using the low-density Affymetrix GeneChip 500 K, and 96.7% using the high-density Illumina Immunochip. For amino acid polymorphisms within HLA genes, we achieve 98.6% and 99.3% accuracy using the Affymetrix GeneChip 500 K and Illumina Immunochip, respectively. Finally, we demonstrate how imputation and association testing at amino acid resolution can facilitate fine-mapping of primary MHC association signals, giving a specific example from type 1 diabetes.