Caitlin Smith has a B.A. in biology from Reed College, a Ph.D. in neuroscience from Yale University, and completed postdoctoral work at the Vollum Institute.
Genetic variation is essential for species to survive changes in their environment. One among many sources of genetic variation, copy-number variation (CNV) is receiving increased attention of late, not only for its prevalence but also for the newly emerging recognition of its role in such human diseases as obesity, heart disease, cancer, autism and schizophrenia.
The consequence of larger-scale genomic events such as deletions, duplications, inversions and translocations, CNV refers to the occurrence of a different or variable number of copies of a segment of DNA, as opposed to single-nucleotide polymorphisms (SNPs), which are differences in single nucleotides. CNVs are more common than many people had thought. “One of the surprising findings is that CNVs and indels [insertions or deletions] encompass over 10 times the total number of nucleotides of genetic variation when compared to SNPs,” says Steve Scherer, director of the Centre for Applied Genomics at the Hospital for Sick Children, University of Toronto.
Traditionally, geneticists identified CNVs using fluorescence in situ hybridization (FISH) and early microarrays, but those methods had limited resolution. Contemporary technologies offer much sharper images of the genome.
Improving resolution
Today’s microarrays little resemble the ones first used for CNV analysis. With smaller, more densely packed spots and shorter, more targeted probes, new microarrays offer higher-resolution genome analyses than ever before, allowing detection of ever-smaller structural variations.
“In the initial discovery studies, we were scanning the genome at 1 Mb resolution, in 2010 we were reaching into the 100 kb range, and now we can get readily down to 10 kb,” says Scherer. “Since the average-sized gene is about 60 kb, this means often the CNVs we are finding occur within a gene, which enables more specificity in genotype and phenotype correlations studies.”
Tools designed to study SNPs are also proving valuable for CNV analysis. Affymetrix’s Genome-Wide Human SNP Array 6.0, for instance, uses more than 1.8 million SNP markers to evaluate copy-number status across the genome in a single experiment. The company’s new CytoScan® Cytogenetics Solution is another type of genome-wide microarray. “It provides coverage of every gene in the genome and is designed to interrogate both nonpolymorphic and polymorphic regions of the genome,” says Dara Wright, vice president of clinical applications marketing at Affymetrix. “Since clinical cytogenetics is very much a learning science, using a targeted array can significantly constrain clinical yield. A whole-genome microarray enables researchers to analyze copy-number variation in a hypothesis-free manner.” Affymetrix’s CytoScan HD Array has more than 2.6 million copy-number markers and can reliably detect 25 kb to 50 kb copy-number changes genome-wide, according to Affymetrix’s website.
For cancer researchers, Affymetrix offers a copy-number assay called OncoScan™ FFPE, which covers the whole genome but has enriched coverage for oncogenes and tumor suppressors. Not only does the OncoScan assay accommodate very small starting sample input (<80 ng DNA, which is a must with precious patient samples), it also targets smaller regions of genomic DNA for binding (about 40 base pairs). “This is critical, as most tumor biopsies are fixed in formalin and stored in paraffin, which can modify and degrade the DNA over time,” says Wright. Offered as a service by Affymetrix today, the OncoScan assay will be available for purchase as a kit later in 2013.
Other manufacturers also offer high-resolution microarrays for CNV analysis, including Illumina’s Infinium HumanCore BeadChip array and Agilent Technologies’ Human Genome CGH Microarrays.
According to Hakon Hakonarson, Associate Professor of Pediatrics at University of Pennsylvania School of Medicine, these high-resolution microarrays enable researchers to detect smaller genomic changes than ever before, revealing a level of genomic variability they didn’t even know existed. “What these tell us is that there are several CNVs that are small and have disease impact. We could not detect them before so did not know about them.”
PCR alternatives
Another approach to CNV detection is PCR. Real-time quantitative PCR (qPCR) and digital PCR (dPCR) both are more quantitative than CNV microarrays (that is, better at measuring fold changes in gene copies), and they can measure CNVs in smaller genomic regions. PCR-based assays also have an advantage in terms of sample throughput relative to microarrays and sequencing approaches.
The difference between dPCR and qPCR stems from the method they use to count a given target. qPCR works by measuring the number of amplification cycles required for a fluorescence signal (representing the target) to surpass some threshold value. That cycle number is then used to calculate the transcript abundance by comparison to a standard curve. In contrast, dPCR counts target molecules directly.
In digital PCR the reaction mixture is initially divided into many (thousands to millions) tiny volumes. These partitions are so tiny that each will generally contain either zero or one DNA molecules. In effect, each reaction becomes binary—either positive or negative. Counting these discrete signals (and applying a Poisson correction to account for wells with more than one copy) enables one to quantify the absolute number of DNA molecules in the sample, explains Jennifer Berman, a senior scientist at Bio-Rad Laboratories’ Digital Biology Center.
“The idea behind digital PCR is that if you partition a sample, you gain increased sensitivity and precision in your measurements that’s well beyond, say, conventional qPCR,” Berman says.
How that partitioning is accomplished varies by system. Bio-Rad’s QX100™ Droplet Digital™ PCR System and RainDance Technologies’ RainDrop™ System both use tiny droplet partitions, while Life Technologies’ QuantStudio™ 3D Digital PCR System uses tiny wells in a dPCR plate.
Fluidigm’s qdPCR 37K™ IFC system, leveraging the company’s strength in microfluidic design, partitions dPCR reactions into tiny microfluidic chambers. According to Ramesh Ramakrishnan, director of molecular biology and assay development at Fluidigm, this system includes a built-in way to monitor reaction efficiency.
“The danger of running end-point assays, which are most of the dPCR systems, is that it diminishes your ability to discriminate true signals from false positives,” Ramakrishnan says. In contrast, Fluidigm’s system also monitors the qPCR reaction continuously so researchers can be assured their amplification enzymes are working equally efficiently for both experimental and reference assays. “While our system does measure in digital mode, it also collects real-time images for every cycle of PCR. It’s a built-in method to find out whether the copy-number results are true, or a false positive, for every one of the spots that you measure in digital mode.” In some situations, such as diagnostic cancer tests, this additional safety check is extremely valuable.
Though not technically dPCR, NanoString Technologies’ nCounter CNV CodeSets also provide high-resolution CNV measurements using a unique system called digital-detection technology, in which molecular “barcodes” comprised of fluorescent tags are detected and counted by single-molecule imaging. This technology has powerful multiplexing capabilities, enabling researchers to assay up to 800 loci in a single tube.
Various qPCR assays for CNV detection are available from such manufacturers as Life Technologies, Fluidigm, and Qiagen, which also recently released a different kind of tool for improving copy-number measurements using qPCR. Normally, determining copy numbers by qPCR requires a reference gene for comparison. Qiagen’s new reference assay relies on a repetitive sequence located at multiple sites in the genome. “This assay improves the accuracy of copy-number measurements by rendering the copy-number determination insensitive to a copy-number change at a single reference locus and providing a better measurement of total DNA input that correlates more closely with other methods,” says Brian McNally, global product manager for Sample & Assay Technologies at Qiagen.
CNVs by NGS
The final strategy for CNV detection is next-generation DNA sequencing (NGS). But there are still some technical hurdles to overcome before this becomes a common approach.
“Exome and whole-genome sequencing are really good at finding small single-letter and insertion or deletion changes—say, up to about 100 bp—but deficiencies in computer programs make it hard to find intermediate and larger CNVs,” says Scherer. A disadvantage of using sequencing for CNV analysis is that it can take days or weeks when factoring in the bioinformatic challenges posed by the results—compared with a few hours for a PCR run. Another caveat is ensuring that sequencing results are real rather than artifactual, which may entail deeper sequencing. But Ramakrishnan says NGS still excels at one important thing—discovering unknowns. He notes that using PCR for an in-depth follow-up on those discoveries is a particularly powerful combination of techniques.
Despite the recent technological advances, scientists still have a lot to learn about CNVs. For example, what is a normal amount of variation? This must be defined to identify something as aberrant, especially when CNV events can be diverse or low-frequency. Wright also notes the importance of associating CNV events “with clinical annotations to strengthen the phenotype-genotype correlation or the genomic structure-disease state correlation,” she says. “In all instances, there is also a desire to collect clinical outcomes information to strengthen the evidence of clinical utility for emerging CNV analysis techniques.”
Image credit: iStockPhoto.com