The Genom of Homo sapiens.pdf

HUMAN MUTATIONAL PROFILING 25Another aspect of DNA sequencing is focused on obtainingthe highest fidelity of sequence assignment, orbase-calling, to the resulting data. Specifically, base-callingalgorithms must be tuned to detect polymorphisms andflag them for downstream analysis. This presents somewhatof a challenge, since polymorphic sites typically exhibittwo unique characteristics that are not recognizedwell by most DNA sequence base-calling algorithms.Namely, a polymorphic site exhibits a significant drop inpeak height relative to neighboring (non-polymorphic)peaks, and contains a smaller underlying peak producedby the second allele. Although the number of commercialsoftware packages aimed at polymorphism detection is increasing,we have used the publicly available polyphredpackage of Nickerson (Nickerson et al. 1997) to identifyand tag polymorphic sites in mutational profiling data.This algorithm has its functions fully integrated within thephred/phrap/consed software packages (Ewing and Green1998; Ewing et al. 1998; Gordon et al. 1998) that we haveused in the production, assembly, and viewing of de novogenome sequence data. As such, we have the ability to tagpolymorphic sites and assemble (via phrap) the componentreads that constitute the exonic (or full gene) sequencesfor a given patient sample. Large-scale comparisonsof cases versus controls or normal versus tumorsamples can then be examined using the consed sequencealignment viewer, in which the polyphred tags can bereadily found and the underlying trace data examined tovalidate the polyphred assignment. Ultimately, these datacan be combined into a visual genotypes output (Nickersonet al. 1998; Rieder et al. 1998) that enables large-scaleexamination and clustering of polymorphic sites for multiplegenes in multiple patient samples.Underlying all of the data collection aspects of a mutationalprofiling project is the crucial aspect of sampletracking, which must be perfect and unequivocal in orderfor results and conclusions to be accurate and meaningful.In our laboratory, we rely on an Oracle database and accompanyingschema to enable barcode-based sampletracking of patient samples and associated data throughoutthe mutational profiling pipeline.MUTATIONAL PROFILING PROJECTSAT WASHINGTON UNIVERSITYPulmonary Surfactant Protein B DeficiencyNewborn respiratory distress syndrome (N-RDS) is themost frequent respiratory cause of death and morbidity ininfants under one year of age in the United States (Guyeret al. 1998). N-RDS has most commonly been attributedin premature infants to pulmonary surfactant deficiencydue to developmental immaturity of surfactant production(Avery and Mead 1959; Whitsett and Stahlman1998), since type II pneumocytes do not appear prior to32–34 weeks of gestation and lack the ability to producefunctional surfactant. Surfactant replacement therapy hasbeen associated with a decline in N-RDS-associated mortality,but not in long-term respiratory morbidity, hencerevealing the contribution of genetic causes of N-RDS ininfancy. Genetic disruption of surfactant protein B geneexpression has provided unambiguous human and murinerespiratory phenotypes in newborn infants and pups. Surfactantprotein B deficiency was the first identified geneticcause of lethal N-RDS in infants and was attributedto a mutation resulting from a 1-bp deletion and a 3-bp insertionat codon 121 of the gene (121ins2) (Nogee et al.1993). This mutation creates a new SfuI restriction site,allowing rapid detection of the mutation, and leads to atranslation stop signal at codon 214. A truncated, unstablesurfactant protein B transcript results, but no proteinis synthesized (Nogee et al. 1994; Beers et al. 2000). Theclinical phenotype of infants homozygous for the 121ins2mutation is consistent: They are born at full-term, developRDS within the first 12–24 hours after birth, haveno associated extrapulmonary organ dysfunction, andwithout lung transplantation will expire within the first1–6 months of life (Hamvas et al. 1997; Nogee 1997).Studies of compound heterozygotes of 121ins2 indicatethat a minimum net production of surfactant protein B between10% and 50% is essential for survival (Ballard etal. 1995; Yusen et al. 1999).The DNA sequence of the surfactant protein B geneand its regulatory regions span 11 kb of the humangenome and are contained in 11 exons, with exons 6 and7 encoding the mature protein and exon 11 providing a3´-untranslated region (Whitsett et al. 1995). The mutationalspectrum of this gene in affecteds has been profiledby previous studies, with the 121ins2 mutation found on~60% of chromosomes (Nogee et al. 2000). In addition tothis predominant mutation, 16 exonic mutations havebeen characterized. Studies of murine knockout or transgeniclineages with disrupted or altered surfactant proteinB gene expression have mimicked many of thehistopathologic features of the disease in humans, suggestingthat mutations resulting in altered processing,folding, intracellular itinerary, or phospholipid interactionmay be clinically significant. Therefore, we have initiateda project to perform mutational profiling across thisgene in multiple patient samples, in an effort to bettercharacterize those mutations in and around the surfactantprotein B gene that appear to predispose infants to RDS.Because of the relatively compact nature of the SP-Bgene, and our desire to understand both coding and noncodingmutational profiles for affecteds in this study, wetook an approach that focused on the entirety of the genesequence, including regions just 5´ of the first exon andthe 3´ UTR, represented as ~500-bp PCR amplicons havinga small amount of sequence overlap with adjoiningproducts. Most of the patient DNAs used in these studiesare obtained from Chelex-based extraction of dried bloodspots taken at birth. The patient samples designated forthis study represent large numbers of patients (~10,000each) from a diverse geographical collection, includingthe state of Missouri; Cape Town, South Africa; Seoul,South Korea; and Oslo, Norway. Comparisons of the mutationalprofiles for these distinct geographic and ethnicgroups, when correlated to phenotype, will enable an estimationof ethnic frequency of the predominant mutationsubtypes, including 121ins2.