HUMAN MUTATION 28(5), 517^521, 2007 MEETING REPORT The 2006 Human Genome Variation Society Scientific Meeting William S. Oetting 1,2Ã 1 School of Pharmacy, University of Minnesota, Minneapolis, Minnesota; 2 Institute of Human Genetics, University of Minnesota, Minneapolis, Minnesota Communicated by Mark Paalman The annual scientific meeting of the Human Genome Variation Society (HGVS) was held on the 9th of October, 2006, in New Orleans, Louisiana. This year’s annual meeting had two main themes, ‘‘Tools to Evaluate Pathogenicity’’ and ‘‘The Human Variome Project.’’ The ability to determine if a DNA variant affects the phenotype is important if we wish to understand the genetic contribution to disease. Genetic variants are continually being identified in research and molecular diagnostic laboratories, but functional tests are not always available. Attempts are now being made to create software that will help us determine if a variation will affect either the function of the protein, the expression of the gene, or the stability and processing of the mRNA. For the second theme, there is an interest in creating a database that brings together genetic variation with phenotypic variation in individuals. The Human Variome Project was created to begin this process. Now that the human genome sequence is all but completed, the next phase of the human genome era will be to associate genetic variation with its effect on the phenotype and differing disease states. At this scientific meeting there were also several papers focusing on the identification, classification, and functional effects of variation. These talks are representative of the questions, problems, and solutions that are being considered by researchers involved in the study of variation in the human genome. Hum Mutat 28(5), 517–521, 2007. r r 2007 Wiley-Liss, Inc. KEY WORDS: gene variation; HGVS; meeting report; Human Variome Project; databases INTRODUCTION The annual meeting of the Human Genome Variation Society (HGVS; www.hgvs.org) was divided into several sections. The two main sessions were ‘‘Tools to Evaluate Pathogenicity and a new initiative ‘‘The Human Variome Project.’’ Other topics included analysis of genetic disorders, methodology for the identification of variants, copy number variants, and microinsertions/microdele- tions. A major goal of the HGVS is to provide a forum for discussing ways to identify and understand the impact of genetic variation on the phenotype and to develop methods to disseminate this information to researchers and clinicians. This annual scientific meeting provides a venue for this discussion. Tools to Evaluate Pathogenicity This session was opened by Shamil R. Sunyaev of the Brigham & Women’s Hospital and the Harvard Medical School, Boston, Massachusetts. He spoke on ‘‘Rare missense polymorphisms: the good, the bad and the ugly.’’ There are, on average, 100 new mutations in each individual. It is therefore important to determine which variants are functional, potentially leading to a change in the phenotype or result in disease. The analysis of systematic deep resequencing datasets together with data on human-chimpanzee sequence divergence and human mutations causing Mendelian diseases suggests that 20% of de novo missense mutations are strongly detrimental, 27% neutral, and 53% mildly deleterious. Surprisingly, allele frequency can serve as a reliable predictor of deleterious amino acid substitutions. More than half of nonsynonymous SNPs (nsSNPs) with an allele frequency below 1% were shown to be mildly deleterious whereas the fraction of deleterious mutations among common polymorphisms is very low. This analysis suggests that multiple low frequency variants may cause common disease phenotypes. In these cases, association studies may have difficulties identifying the genetic variation underlying these diseases. Instead, association studies aimed at the detection of rare missense mutation enrichment may allow for the identification of the true disease-causing variants. If one accepts this premise, the identification of functional mutations becomes increasingly important. PolyPhen (http://genetics.bwh.harvard. edu/pph) can be used to predict the functional effect of nsSNPs. PolyPhen predicts the effect of nsSNPs from comparative sequence analysis and from protein 3D-structure and annotation when available. A new version of PolyPhen improves prediction accuracy mostly due to a new sequence alignment pipeline. The pipeline consists of homology searches (BLAST; http://www.ncbi.nlm.nih. gov/blast), constructing multiple alignments (MUSCLE; http://www.drive5.com/muscle), a filtering algorithm (LEON), a clustering algorithm (SECATOR; http://www-bio3d-igbmc. u-strasbg.fr/ ~ wicker/secator/secator.html) and profile scores (PSIC; http://strand.imb.ac.ru/psic) to create a predictive score. Published online 13 March 2007 in Wiley InterScience (www. interscience.wiley.com). DOI 10.1002/humu.20489 Received 11 December 2006; accepted revised manuscript 21 December 2006. Ã Correspondence to:William S. Oetting, Ph.D., Institute of Human Genetics, MMC 485, 420 Delaware St. S.E., University of Minnesota, Minneapolis, MN 55455. E-mail: bill@lenti.med.umn.edu r r 2007 WILEY-LISS, INC.