Summary of webinar “CNVs vs. SNPs: Understanding Human Structural Variation in Disease”

NHGRI CNV image Do you still believe that monozygotic, or identical, twins online canadian pharmacy are really genetically identical?

Or that we are all 99.9% genetically similar to each other? Well I certainly did, and boy was I wrong!

It turns out that CNVs (Copy Number Variations) are causing the “facts” some of us learned in Molecular Biology 101 to be rewritten. If you, like me, thought that what you learned years ago was still true, then there is a great webinar you may want to watch. It is brought to you by Science/AAAS, and it features three prominent experts in genetic variability, Drs. Charles Lee, Lars Feuk and Alexandra Blakemore.

The moderator is Dr. Sean Sanders, who is the Commercial Editor of Science. Even those of you that are up to speed on the current research can find many interesting facts and learn about the new techniques used to study CNVs, or just genetic variability in general. It turns out that CNVs are much more prevalent than was previously thought. You hear so much about SNPs that it seems like they are the source of genetic variability that we should be most concerned about, but CNVs are catching up real fast. This new field is rapidly advancing because of major technology breakthroughs.

All of the panelists present a short talk highlighting the prevalence, importance and experimental limitations of studying CNVs and their role in normal human variability, as well as in disease. They present some of their own data and discuss the future direction of this young field. This is followed by a very interesting question and answer session where they allowed listeners to email their questions. It may even turn out that CNVs are the reason that your personality, IQ, height and weight differ from your colleagues, friends and family. So not only is this an exciting new field, but it is certainly one we can all relate to!

The first panelist, Dr. Charles Lee, provided a nice general background, history of the field and current statistics. CNVs can be either gains or losses of genomic segments, and it wasn’t until 2004 that researchers realized SNPs were not the only important contributors to genetic variability. Structural genomic variants are defined as non-SNPs and within this group CNVs seem to compose the largest fraction. It is believed that somewhere been 5-25% of the human genome is copy number variable and that each person may have about 1500 inherited CNVs with an average size of 20kb. The mutation rate, or the number of CNVs that have arisen per generation, is still totally unknown. However, it is known that CNVs are generally located outside of genes and ultra-conserved elements. Obviously that doesn’t mean that they may not still affect a gene expressed even far away, though. For those CNVs that overlap genes there is a very interesting trend. They appear to be in genes that are responsible for our interaction with the environment. For example, they cluster in GO categories like sensory perception/neurophysiological processes, immune response/inflammation, drug detoxification and more. Dr. Lee also presents some of the molecular mechanisms by which CNVs may have arisen and highlights evidence of the roles of CNVs in diseases like Crohn’s and psoriasis. Most interesting here is the fact that copy number increases or decreases can both be deleterious. Lastly, he reviews and compares the current genome wide technologies available to study CNVs and discusses why normal individuals are no longer considered to be 99.9% genetically identical.

NHGRI CNV image3Next, Dr. Lars Feuk highlighted how important it is to produce a high resolution map of CNVs. Currently, the map is still too crude and that presents many experimental hurdles. Dr. Feuk is the curator of the Database of Genomic Variants which houses these data. He explains some of the technological reasons why CNV data lags behind SNPs and large cytogenetic genomic data. The major reason is because CNVs are mid-range in size, generally falling between the range of 100bps and 20kb. Large changes, like chromosomal rearrangements have been picked up cytogenetically for many years and SNPs have been detected by sequencing, but the techniques for accurately detecting these mid-size range changes are not as good. He reviews how different platforms can give you very different results. This can make disease associations very difficult. You can get both plate and batch effects and results vary depending on input DNA quality and analysis tools used. Extensive validation of the data is a crucial part of these studies. In the future Dr. Feuk believes that next generation sequencing approaches will be very important in solving the experimental obstacles and that soon we will no longer be discussing SNPs and CNVs-only genetic variation.

Dr. Blakemore focused on the correlation of CNVs with phenotype. The take home message here was how very hard it is to assign phenotypes to CNVs currently, because if you look at “normal” individuals you find many CNVs. And their importance appears to have nothing to do with their size as previously thought. You can have big CNVs in normal individuals and small ones that are associated with disease. Copy number also doesn’t seem to have any clear cut correlation.  She also presents an estimate that about 26% of CNVs include exonic sequences and shows that the genes affected cluster into groups involved in environmental response, similar to Dr. Lee. The fact that monozygotic twins have CNVs that make them not genetically identical is discussed and the question of how these arise is raised. It is possible they are simply accumulating during aging. The incidence of these de novo CNVs is not yet established, but it looks to be around 1-3%.  Here you can also learn that there seems to be subclasses of CNVs-those that are stable and old and would be in strong linkage disequilibrium with SNPs and those that are new. This is one of the many problems in predicting CNVs from the existing data. There are many algorithms being tested for their predictive abilities, but it appears from this discussion that there is no consensus as to which one is the best yet.

The question and answer session was also very enlightening and enjoyable. My favorite question, which was asked to all of the panelists, was to describe the biggest challenge they see. Although some parts of their answers varied slightly, there was certainly agreement that the most necessary and important next step is to build a high resolution map of CNVs to use as a tool to design further studies.

For any of you interested in learning more or viewing some of these data make sure to check out the Database of Genomic Variants. And you can also read a great blog post on CNVs and autism here. In addition, an exceptional paper by Kidd et al., entitled “Mapping and sequencing of structural variation from eight human genomes” appeared recently in Nature (2008 May 1;453(7191):56-64).  The most important points are highlighted for you to read in a summary post by Mary or you can download the paper here. Agilent provides even more CNV information, such as interviews and article reviews, and a genomics newsletter that you can sign up for at no cost.
Jeffrey M. Kidd, Gregory M. Cooper, William F. Donahue, Hillary S. Hayden, Nick Sampas, Tina Graves, Nancy Hansen, Brian Teague, Can Alkan, Francesca Antonacci, Eric Haugen, Troy Zerr, N. Alice Yamada, Peter Tsang, Tera L. Newman, Eray Tüzün, Ze Cheng, Heather M. Ebling, Nadeem Tusneem, Robert David, Will Gillett, Karen A. Phelps, Molly Weaver, David Saranga, Adrianne Brand, Wei Tao, Erik Gustafson, Kevin McKernan, Lin Chen, Maika Malig, Joshua D. Smith, Joshua M. Korn, Steven A. McCarroll, David A. Altshuler, Daniel A. Peiffer, Michael Dorschner, John Stamatoyannopoulos, David Schwartz, Deborah A. Nickerson, James C. Mullikin, Richard K. Wilson, Laurakay Bruhn, Maynard V. Olson, Rajinder Kaul, Douglas R. Smith, Evan E. Eichler (2008). Mapping and sequencing of structural variation from eight human genomes Nature, 453 (7191), 56-64 DOI: 10.1038/nature06862