A recent study by the Wellcome Trust Case Control Consortium has found that common copy number variants assessed by current array technology have little influence on common human diseases. In response to the findings, the researchers vowed to continue using CNV-typing array platforms in their research, while focusing more on rarer variants.
In the study, published in Nature online last week, the WTCCC described how it used a custom-designed Agilent Technologies comparative genomic hybridization array to assess common CNVs in about 16,000 individuals diagnosed with eight common diseases, and 3,000 healthy controls.
However, the effort turned up only three genetic loci that were associated with disease and were confirmed by follow-up experiments. This finding suggests that common CNVs may not be the sole cause of these or other common diseases.
WTCCC chair Peter Donnelly, who also is director of the Wellcome Trust Centre for Human Genetics and a statistics researcher at the University of Oxford, said in a statement that a "strong view" existed in the research community that CNVs would be "important for common disease, and that they would explain much of the missing heritability." Instead, the WTCCC no longer believes this is the case, and Donnelly described the consortium's results as "surprising and disappointing for some parts of the community."
Matthew Hurles, a human genetics researcher at the Wellcome Trust Sanger Institute and co-author on the paper, said in a statement that it is "unlikely that common CNVs play a major role in the genetic basis of common diseases, either through particular CNVs having a strong effect or through a large number of CNVs each contributing a small effect.
"This is certainly the case for the diseases that we studied, but is likely to be the case for other common diseases, too," Hurles added. Neither Donnelly nor Hurles responded to e-mails seeking comment.
The study is certain to gain the attention of big array manufacturers, all of which have looked to make CNV content available to the market in recent years. Agilent, for instance, last year launched three catalog CNV arrays, including one designed by the WTCCC, to allow customers to integrate CNV analysis into their genome-wide association studies (BAN 4/7/2009). Roche NimbleGen similarly launched two CNV-themed chips last year (BAN 9/29/2009), while Affymetrix and Illumina have sought to make more CNV content available on their next-generation SNP genotyping chips.
According to some authors of the paper, it is unlikely the results could hurt array companies that make CNV chips. Rather, they believe the Nature results will guide the development of the next generation of CNV arrays.
Co-author Stephen Scherer, director of the Centre for Applied Genomics at Toronto's Hospital for Sick Children, called the WTCCC study a "first-pass analysis" to test for common trends in the eight diseases being tested.
"While there was no glaringly obvious capture of the so-called 'missing heritability' in common CNVs, this does not mean the same approach will not be more successful for other diseases," Scherer told BioArray News this week.
Scherer is also an author of a second paper that appeared in Nature last week, where WTCCC researchers used tiling oligonucleotide microarrays composed of 42 million probes to generate a comprehensive map of 11,700 copy number variations. In that paper, the authors determined that "for complex traits, the heritability void left by genome-wide association studies will not be accounted for by common CNVs."
According to Scherer, both papers include "some caveats to the approaches used that still need to be addressed." For instance, the authors acknowledged in the first paper that the findings should be "interpreted within the context of several limitations."
The WTCCC authors admitted that for some CNVs they could not reliably assign copy-number classes. The authors estimated that "just under half of these were not polymorphic in our data, being either false positives in the discovery experiment, or very rare in the UK population" and that for the remainder, they were also "unable to perform reliable association analyses based directly on intensity measurements." Such CNVs might be "systematically different from those that we do type successfully, in which case it is not possible to extrapolate from our results to their potential role in human disease," they wrote.
The authors of the first paper also noted that their experiment was "powered to detect associations with common copy number variation and our observations and conclusions do not necessarily generalize to the study of rare copy number variants."
Charles Lee, a co-author on both papers and the director of cytogenetics at Harvard Cancer Center, similarly argued that the limitations of the WTCCC study should be taken into account when considering its implications for future research.
In particular, Lee told BioArray News this week that the study's conclusions "reflected genotypable common CNV data" and excluded "common CNVs that were difficult to genotype on the array platforms used."
He added that the WTCCC study excluded multiallelic and complex CNVs, which are difficult to accurately genotype with the array platform used in the study. "Interestingly, many CNVs that are currently associated with human phenotypes are of a multiallelic or complex nature," he said.
[ pagebreak ]
Lee said that such findings should encourage array manufacturers to make "new CNV typing array platforms and complementary technologies that do a better job genotyping these particular CNV regions."
CNV-focused arrays are still the "most cost-effective technology for genome-wide association studies," Lee said, but future CNV-focused array platforms could be improved to "include more CNVs of lower minor allele frequencies and using CNV data from other world populations, empirically enhance probe selection for a given CNV region to improve on genotypability, and perhaps add probes targeting CNV breakpoints."
The recent WTCCC paper indicated that it is "unlikely that common CNVs will account for the vast majority of the missing heritability for common diseases, but at the same token it does not discount the idea that common CNVs could account for a smaller yet significant amount of the missing heritability," Lee said. "I am optimistic that the latter will be the case."
Scherer similarly said that the papers were likely to lead to the creation of new chips. "I do not [think] these papers will necessarily negate any companies' business plans, but instead perhaps advise on the most appropriate technologies to further test this hypothesis going forward," he said. "The data also lend the growing support for rare variants in some common diseases."
A New Batch of CNV Chips
Rohaizah James, product manager of CGH and CNV arrays at Roche NimbleGen, said the firm is well aware of the WTCCC study, and is already planning to launch new chips with more rare variant content later this year.
"We have realized the importance of rare CNVs in determining causative factors of disease, which is why Roche NimbleGen is continuing to expand our CNV portfolio with the next generation of CNV arrays," James told BioArray News this week.
The new arrays, which will be available in the second half of this year, will offer "unprecedented probe density to allow for the highest-resolution array detection of these rare variants while also including common CNVs to provide researchers with the full CNV picture," James said.
She noted that "while researchers now believe that rare CNVs are likely to be the causative factors for disease, identifying common CNVs is still important in understanding the entire picture of what the genome holds." Arrays, James added, will "continue to be an important tool in CNV research due to the parallel processing of samples and their cost-effectiveness over other genomic technologies."
Agilent did not respond to questions in time for this publication. Still, the firm has in the past said that it will likely develop new CNV-themed arrays in addition to the three catalog chips it already sells. Dione Bailey, Agilent's product manager of CGH and CNV microarrays, told BioArray News in August 2009 that it "plans to launch additional arrays over the next 12 to 16 months as we expand our platform" (BAN 9/29/2009).