Linkage disequilibrium maps constructed with common SNPs are useful for first-pass disease association screens.

To develop an efficient strategy for mapping genetic factors associated with common diseases, we constructed linkage disequilibrium (LD) maps of human chromosomes 5, 7, 17, and X. These maps consist of common single nucleotide polymorphisms at an average intermarker distance of 100 kb. The genotype data from these markers in a panel of American samples of European descent were analyzed to produce blocks of markers in strong pair-wise LD. Power calculations were used to guide block definitions and predicted that high-level LD maps would be useful in initial genome scans for susceptibility alleles in case-control association studies of complex diseases. As anticipated, LD blocks on the X chromosome were larger and covered more of the chromosome than those found on the autosomes.

Allelic association with SNPs: metrics, populations, and the linkage disequilibrium map.

Comparison of different metrics, using three large samples of haplotypes from different populations, demonstrates that rho is the most efficient measure of association between pairs of single nucleotide polymorphisms (SNPs). Pairwise data can be modeled, using composite likelihood, to describe the decline in linkage disequilibrium with distance (the Malecot model). The evidence from more isolated populations (Finland, Sardinia) suggests that linkage disequilibrium extends to 427-893 kb but, even in samples representative of large heterogeneous populations, such as CEPH, the extent is 385 kb or greater. This suggests that isolated populations are not essential for linkage disequilibrium mapping of common diseases with SNPs. The in parameter of the Malecot model (recombination and time), evaluated at each SNP, indicates regions of the genome with extensive and less extensive disequilibrium (low and high values of in respectively). When plotted against the physical map, the regions with extensive and less extensive linkage disequilibrium may correspond to recombination cold and hot spots. This is discussed in relation to the Xq25 cytogenetic band and the HFE gene region.

The optimal measure of allelic association.

Allelic association between pairs of loci is derived in terms of the association probability rho as a function of recombination theta, effective population size N, linear systematic pressure v, and time t, predicting both rho(rt), the decrease of association from founders and rho(ct), the increase by genetic drift, with rho(t) = rho(rt) + rho(ct). These results conform to the Malecot equation, with time replaced by distance on the genetic map, or on the physical map if recombination in the region is uniform. Earlier evidence suggested that rho is less sensitive to variations in marker allele frequencies than alternative metrics for which there is no probability theory. This robustness is confirmed for six alternatives in eight samples. In none of these 48 tests was the residual variance as small as for rho. Overall, efficiency was less than 80% for all alternatives, and less than 30% for two of them. Efficiency of alternatives did not increase when information was estimated simultaneously. The swept radius within which substantial values of rho are conserved lies between 385 and 893 kb, but deviation of parameters between measures is enormously significant. The large effort now being devoted to allelic association has little value unless the rho metric with the strongest theoretical basis and least sensitivity to marker allele frequencies is used for mapping of marker association and localization of disease loci.

Regions of low single-nucleotide polymorphism incidence in human and orangutan xq: deserts and recent coalescences.

While scanning for single-nucleotide polymorphisms (SNPs) in the human Xq25-q28 region of CEPH families, we found six long "deserts" of low SNP incidence representing 28% of the investigated genome. One was 1.66 Mb in length. To determine whether these SNP deserts were due to reduced input of mutations or to recent coalescent events such as bottlenecks or selective sweeps, comparative sequence was determined from a female orangutan. The mean divergence was 2.9% and was not reduced in deserts compared with nondesert regions. Thus, the best explanation for the SNP deserts is recent coalescent events in humans. These events are the cause of substantial variation in human noncoding SNP incidence. In addition, the mutational spectrum in humans and orangutans was estimated as 63% AG (and CT), 17% AC (and GT), 8% CG, 4% AT, and 8% insertion/deletions. The average lifetime of a SNP destined to become fixed for a new allele between these species was estimated as 284,000 years.

Juxtaposed regions of extensive and minimal linkage disequilibrium in human Xq25 and Xq28.

Linkage disequilibrium (LD), or the non-random association of alleles, is poorly understood in the human genome. Population genetic theory suggests that LD is determined by the age of the markers, population history, recombination rate, selection and genetic drift. Despite the uncertainties in determining the relative contributions of these factors, some groups have argued that LD is a simple function of distance between markers. Disease-gene mapping studies and a simulation study gave differing predictions on the degree of LD in isolated and general populations. In view of the discrepancies between theory and experimental observations, we constructed a high-density SNP map of the Xq25-Xq28 region and analysed the male genotypes and haplotypes across this region for LD in three populations. The populations included an outbred European sample (CEPH males) and isolated population samples from Finland and Sardinia. We found two extended regions of strong LD bracketed by regions with no evidence for LD in all three samples. Haplotype analysis showed a paucity of haplotypes in regions of strong LD. Our results suggest that, in this region of the X chromosome, LD is not a monotonic function of the distance between markers, but is more a property of the particular location in the human genome.

A high-density single-nucleotide polymorphism map of Xq25-q28.

A high-density single-nucleotide polymorphism (SNP) map was developed for Xq25-q28 using a targeted approach to SNP discovery. This high-density map includes 217 new SNP markers, and 117 are informative in the CEPH parent population with >20% minor allele frequency. The average distance between SNP markers is 100 kb in the targeted regions. This is the densest genetic map of Xq25-q28 to date. The SNP markers are presented in order by their distance in megabases along the X chromosome, and the markers from the current genetic map are placed using the same scale to produce an integrated map of the region.

Efficient approach to unique single-nucleotide polymorphism discovery.

Single-nucleotide polymorphisms (SNPs) are the most frequently found DNA sequence variations in the human genome. It has been argued that a dense set of SNP markers can be used to identify genetic factors associated with complex disease traits. Because all high-throughput genotyping methods require precise sequence knowledge of the SNPs, any SNP discovery approach must involve both the determination of DNA sequence and allele frequencies. Furthermore, high-throughput genotyping also requires a genomic DNA amplification step, making it necessary to develop sequence-tagged sites (STSs) that amplify only the DNA fragment containing the SNP and nothing else from the rest of the genome. In this report, we demonstrate the utility of a SNP-screening approach that yields the DNA sequence and allele frequency information while screening out duplications with minimal cost and effort. Our approach is based on the use of a homozygous complete hydatidiform mole (CHM) as the reference. With this homozygous reference, one can identify and estimate the allele frequencies of common SNPs with a pooled DNA-sequencing approach (rather than having to sequence numerous individuals as is commonly done). More importantly, the CHM reference is preferable to a single individual reference because it reveals readily any duplicated regions of the genome amplified by the PCR assay before the duplicated sequences are found in GenBank. This approach reduces the cost of SNP discovery by 60% and eliminates the costly development of SNP markers that cannot be amplified uniquely from the genome.

Overlapping genomic sequences: a treasure trove of single-nucleotide polymorphisms.

An efficient strategy to develop a dense set of single-nucleotide polymorphism (SNP) markers is to take advantage of the human genome sequencing effort currently under way. Our approach is based on the fact that bacterial artificial chromosomes (BACs) and P1-based artificial chromosomes (PACs) used in long-range sequencing projects come from diploid libraries. If the overlapping clones sequenced are from different lineages, one is comparing the sequences from 2 homologous chromosomes in the overlapping region. We have analyzed in detail every SNP identified while sequencing three sets of overlapping clones found on chromosome 5p15.2, 7q21-7q22, and 13q12-13q13. In the 200.6 kb of DNA sequence analyzed in these overlaps, 153 SNPs were identified. Computer analysis for repetitive elements and suitability for STS development yielded 44 STSs containing 68 SNPs for further study. All 68 SNPs were confirmed to be present in at least one of the three (Caucasian, African-American, Hispanic) populations studied. Furthermore, 42 of the SNPs tested (62%) were informative in at least one population, 32 (47%) were informative in two or more populations, and 23 (34%) were informative in all three populations. These results clearly indicate that developing SNP markers from overlapping genomic sequence is highly efficient and cost effective, requiring only the two simple steps of developing STSs around the known SNPs and characterizing them in the appropriate populations.

The homozygous complete hydatidiform mole: a unique resource for genome studies.

The most frequent type of complete hydatidiform mole is a 46, XX homozygote formed by the fertilization of an empty ovum by a single haploid sperm that later duplicates its chromosomes to give a diploid tumor. The homozygous nature of these complete hydatidiform moles makes them unique resources for human genome studies. They can serve as homozygous controls in the development of single nucleotide polymorphism (SNP) markers and provide a way to obtain long-range haplotypes that are useful in population studies. The use of a homozygous control makes it possible to estimate the allele frequencies of the SNP markers in any population by sequencing pooled DNA samples. In this report, we present evidence of homozygosity of a complete hydatidiform mole using 20 diallelic markers distributed across the genome. Furthermore, its usefulness as a homozygous control in SNP development and as a resource for long-range haplotype determination is demonstrated using 11 newly discovered loci in the BRCA2 region on chromosome 13q12-q13.

Multiple genetic loci within 11p15 defined by Beckwith-Wiedemann syndrome rearrangement breakpoints and subchromosomal transferable fragments.

Beckwith-Wiedemann syndrome (BWS) involves fetal overgrowth and predisposition to a wide variety of embryonal tumors of childhood. We have previously found that BWS is genetically linked to 11p15 and that this same band shows loss of heterozygosity in the types of tumors to which children with BWS are susceptible. However, 11p15 contains > 20 megabases, and therefore, the BWS and tumor suppressor genes could be distinct. To determine the precise physical relationship between these loci, we isolated yeast artificial chromosomes, and cosmid libraries from them, within the region of loss of heterozygosity in embryonal tumors. Five germ-line balanced chromosomal rearrangement breakpoint sites from BWS patients, as well as a balanced chromosomal translocation breakpoint from a rhabdoid tumor, were isolated within a 295- to 320-kb cluster defined by a complete cosmid contig crossing these breakpoints. This breakpoint cluster terminated approximately 100 kb centromeric to the imprinted gene IGF2 and 100 kb telomeric to p57KIP2, an inhibitor of cyclin-dependent kinases, and was located within subchromosomal transferable fragments that suppressed the growth of embryonal tumor cells in genetic complementation experiments. We have identified 11 transcribed sequences in this BWS/tumor suppressor coincident region, one of which corresponded to p57KIP2. However, three additional BWS breakpoints were > 4 megabases centromeric to the other five breakpoints and were excluded from the tumor suppressor region defined by subchromosomal transferable fragments. Thus, multiple genetic loci define BWS and tumor suppression on 11p15.

The human factor H-related gene 2 (FHR2): structure and linkage to the coagulation factor XIIIb gene.

The human factor H-related gene 2 (FHR2) encodes a serum protein structurally and immunologically related to complement factor H. We describe the isolation and genomic organization of the human FHR2 gene from a yeast artificial chromosome library. The FHR2 gene is organized in five exons and spans about 7 kilobases (kb) of human genomic DNA. A comparison with the corresponding cDNA sequence (clone DDESK59) shows that the analyzed FHR2 gene has a deleted region within exon 4. A new splice acceptor site created in the truncated exon indicates that the analyzed gene could be translated to a truncated protein. Further, we demonstrate that the genes for FHR2 and beta subunit of coagulation factor XIII are located in the same 165 kb YAC DNA. Thus, the three structurally related genes FXIIIb, FHR2, and factor H are linked on human chromosome 1 in the regulators of complement activation (RCA) gene cluster. The physical linkage of the FHR2 and the factor H genes provides additional evidence for a close relatedness of complement factor H and the factor H-related proteins. The linkage and the almost exclusive organization in short consensus repeat-containing domains indicates a close evolutionary relationship of the FXIIIb, FHR2, and factor H genes.

Structure and expression of the human gene for the alpha subunit of prolyl 4-hydroxylase. The two alternatively spliced types of mRNA correspond to two homologous exons the sequences of which are expressed in a variety of tissues.

Prolyl 4-hydroxylase, an alpha 2 beta 2 tetramer, plays a central role in collagen synthesis as it catalyzes the formation of 4-hydroxyproline residues by the hydroxylation of proline in X-Pro-Gly sequences. We report here that the human gene for the catalytically important alpha subunit is more than 69 kilobase pairs and consists of 16 exons. The exons that encode solely protein sequences vary from 54 to 240 base pairs (bp), and the introns vary from 750 to more than 16,000 bp. The 133 bp of 5'-untranslated sequences of the mRNA are coded by two exons, and these sequences contain inverted repeats with a potential for stem-loop formation, which may be involved in translational control of the expression of this gene. The 5'-flanking region contains a TATa motif at -29 relative to the major transcription site but no CCAAT motif. The 5'-flanking region and the downstream sequences contain several motifs that may act as binding sites for various transcription factors. Evidence has previously been reported for a mutually exclusive alternative splicing of RNA transcripts of this gene. The present data indicate that the mutually exclusive sequences found in the mRNAs are coded by two consecutive, homologous 71-bp exons 9 and 10. These exons are identical in their first 5 bp and the overall identity between them is 61% at the nucleotide level and 58% at the level of the coded amino acids. Both types of mRNA were found to be expressed in all of the tissues studied, but in some tissues the type coding for exon 9 or 10 sequences was more abundant than the other type.

Cloning and chromosomal localization of the gene coding for human protein kinase CK1.

A cDNA clone coding for human protein kinase CK1 (casein kinase 1) has been isolated and sequenced demonstrating that it corresponds to a homolog of the CK1 alpha form found in bovine brain. The derived amino acid sequence of the human CK1 alpha is identical to the bovine counterpart except that it contains 12 extra amino acids at the carboxyl end. Using this cDNA sequence and PCR amplification, YAC genomic clones that contain this human CK1 alpha sequence have been isolated. These YACs have been used for fluorescent in situ hybridization in order to localize the human CK1 alpha gene to chromosome 13q13.

2.6 Mb YAC contig of the human X inactivation center region in Xq13: physical linkage of the RPS4X, PHKA1, XIST and DXS128E genes.

X chromosome inactivation is a mechanism of dosage compensation that regulates the expression of mammalian X-linked genes between XY males and XX females. This phenomenon is cis-acting, clonally heritable, and requires the presence of an X inactivation center (XIC). In our attempts to characterize this phenomenon, we have focused on the physical organization of the human XIC localized to Xq13. From previous studies, we had determined that the candidate XIC interval contained two loci (DXS128 and XIST) and was bound by the breakpoints of two structurally abnormal inactivated X chromosomes, a t(X;14) and an idic(Xp). Here we present a refined mapping of the XIC-containing region using the breakpoint of a late replicating rearranged X (rea(X)), and the initial characterization of a set of 40 yeast artificial chromosomes (YACs) derived from the XIC-containing region. These YACs form a 2.6 Mb contig which completely covers the XIC, and physically links the RPS4X, PHKA1, XIST, and DXS128E genes, as well as a laminin receptor pseudogene (LAMRP4). Furthermore, we have determined the relative orientations of these four genes, and have derived a restriction map of the region using the rare cutter enzymes BssHII, EagI, MluI, NruI, SalI, SfiI, SstII (or SacII), and NotI. We have identified at least 9 CpG-rich islands within this region, and have discovered a large (approximately 125 kb) inverted duplication proximal to the XIC based on symmetrical restriction patterns and homologous probes. We estimate the maximum size of the XIC-containing interval to be between 680 kb and 1200 kb, based on the localization of the breakpoints of the rearranged X chromosomes mentioned above. This lays the groundwork for the further characterization of the XIC region and the isolation of other expressed sequences therefrom.

YAC-assisted cloning of transcribed sequences from the human chromosome 3p21 region.

The region surrounding the ZNF35 zinc finger protein gene on 3p21 is of particular interest, as this region of chromosome 3 is frequently involved in rearrangements and/or deletions associated with various human tumors including lung and renal carcinoma. We have analyzed yeast artificial chromosomes (YACs), identified by PCR screening, using oligonucleotides derived from the ZNF35 gene. PFGE and Southern blot/hybridization analysis revealed that the clones cover 750-kb including the ZNF35 gene. The use of specific somatic cell hybrids have allowed us to locate the YAC contig telomeric to the D3F15S2 locus, in a region which is frequently deleted in lung carcinomas. In addition, we have developed a novel cDNA hybridization protocol allowing the isolation of transcribed sequences present in the overlapping YAC clones. Using the cDNA hybridization selection, we have isolated and characterized one transcribed sequence (D3S1362E) from the 3p21 YAC contig and the corresponding cDNA has been isolated. DNA sequencing analysis indicated that the D3S1363E cDNA codes for a putative transcription factor. Northern blot analysis indicated that the D3S1362E sequence hybridized to multiple transcripts in skeletal muscle, and weakly hybridizing transcripts of similar sizes were detected in other tissues.

PCR buffer optimization with uniform temperature regimen to facilitate automation.

To facilitate PCR(1,2) reactions in large numbers with uniform conditions, the annealing temperature was fixed and the stringency of the reactions was adjusted by optimizing the ion conditions of the reaction. The buffer system is based primarily on Tris (T), ammonium (N), and potassium (K) to adapt assay conditions to different primer pairs. The TNK buffers have permitted successful screening of a 60,000-clone yeast artificial chromosome (YAC) library with more than 200 primer pairs.

Chromosomal reassignment: YACs containing both YES1 and thymidylate synthase map to the short arm of chromosome 18.

The YES1 proto-oncogene was mapped previously to human chromosome band 18q21.3 by using isotopic in situ hybridization. Using yeast artificial chromosomes (YACs) as probes and fluorescence in situ hybridization, a strong signal was detected in the region corresponding to 18p11.3. Restriction digests confirmed that the YACs contained the YES1 gene and not other cross-hybridizing, protein-tyrosine kinases. In addition, these YACs were found to contain another 18p11.32 gene, thymidylate synthase. These genes were less than 50 kb apart. Collectively, these data suggest that YES1 maps to 18p11.32 rather than to 18q21.3.

Isolation and structural analysis of a 1.2-megabase N-myc amplicon from a human neuroblastoma.

Oncogene amplification is observed frequently in human cancers, but little is known about the mechanism of gene amplification or the structure of amplified DNA in tumor cells. We have studied the N-myc amplified domain from a representative neuroblastoma cell line, SMS-KAN, and compared the map of the amplicon in this cell line with that seen in normal DNA. The SMS-KAN cell line DNA was cloned into yeast artificial chromosomes (YACs), and clones were identified by screening the YAC library with amplified DNA probes that were obtained previously (B. Zehnbauer, D. Small, G. M. Brodeur, R. Seeger, and B. Vogelstein, Mol. Cell. Biol. 8:522-530, 1988). In addition, YAC clones corresponding to the normal N-myc locus on chromosome 2 were obtained by screening two normal human YAC libraries with these probes, and the restriction maps of the two sets of overlapping YACs were compared. Our results suggest that the amplified domain in this cell line is a approximately 1.2-Mb circular molecule with a head-to-tail configuration, and the physical map of the normal N-myc locus generally is conserved in the amplicon. These results provide a physical map of the amplified domain of a neuroblastoma cell line that has de novo amplification of an oncogene. The head-to-tail organization, the general conservation of the normal physical map in the amplicon, and the extrachromosomal location of the amplified DNA are most consistent with the episome formation-plus-segregation mechanism of gene amplification in these tumors.

A yeast artificial chromosome contig encompassing the type 1 neurofibromatosis gene.

The yeast artificial chromosome (YAC) system (Burke et al., 1987, Science 236: 806-812) allows the direct cloning of large regions of the genome. A YAC contig map of approximately 700 kb encompassing the region surrounding the type 1 neurofibromatosis (NF1) locus on 17q11.2 has been constructed. A single YAC containing the entire NF1 locus has been constructed by homologous recombination in yeast. In the process of contig construction a novel method of YAC end rescue has been developed by YAC circularization in yeast and plasmid rescue in bacteria. YACs containing homology to the NF1 region but mapping to another chromosome have also been discovered. Sequences of portions of the homologous locus indicate that this other locus is a nonprocessed pseudogene.