Finding the origin and impact of altered gene doses

[nsane.forums]

The authors found the precise sequence at CNV breakpoints by identifying DNA in the area, and using it to isolate the right sequences.

In theory, we each have two copies of every gene; one on the chromosome from mom and the other on the chromosome from dad. In reality, this is not necessarily the case. The Human Genome Project revealed that Copy Number Variations (CNVs)—the presence of something other than two copies of a region of DNA due to its duplication or deletion—is quite widespread in humans. CNVs have been shown to be associated with a number of human maladies, including cancer and susceptibility or resistance to HIV infection. Although many CNVs have been identified, it has been difficult to determine the mechanism responsible for producing them. This is partially because fewer than 10 percent of known CNVs have been mapped to base pair resolution, so their precise boundaries are unknown.

High throughput approaches have thus far not been well suited to sequencing these breakpoints. Genome wide shotgun sequencing involves breaking the DNA into short stretches, which also generates breakpoints; designing primers for PCR based sequencing is tough when you don’t know the sequence flanking the region of interest.

Array-based oligonucleotide hybridization to the rescue! Arrays were made using DNA from the area near the breakpoint region, and DNA fragments containing the breakpoints—both test fragments with the potential CNV from three unrelated individuals and reference fragments with one copy of the region—were captured by hybridization. These fragments were then isolated and sequenced. Fragments without CNV breakpoints show contiguous homology to the reference, while those with breakpoints are split, with partial homology on each side of the CNV.

Two types of DNA break repair can account for CNVs, and defining the breakpoint should help in assigning the pathway responsible for a specific event. Non homologous end joining is often associated with tiny regions (1-4bp) of DNA homology, and may also entail the addition of short, nontemplated sequences. Homology directed repair, including nonallelic homologous recombination and single strand annealing, is thought to require longer stretches (>200bp) of homology.

In a recent report in Nature, 1067 CNVs were targeted using a capture array, 324 breakpoints were found, and 315 of these were sequenced (all from deletions, no duplications). The array allowed more CNV breakpoints to be sequenced in this one experiment than the total that had been reported to date. Of these, 105 (33 percent) had 1-367 bp of inserted sequence at the deletion site, while 219 (70 percent) had 1-30 homologous bases at the ends. Only 32 of the deletions containing inserted sequences are flanked by microhomologies; this is a smaller percentage than would be expected if these two repair mechanisms acted independently. Twenty-five breakpoints were blunt ends, with no homology or inserted sequence, suggesting that CNVs do not occur randomly throughout the genome. Together, this data implies that more complex genetic rearrangements than had previously been considered are at play.

Precisely mapping these breakpoints is important not only for figuring out exactly how the CNVs were generated, but also for finding which sequences have been disrupted. This is essential in assessing any functional impact on nearby genes. Of course, these CNVs were taken from somatic cells, and mutational processes acting in germline cells may be different. Yet the authors hope that this is a first step towards the ability to reconstruct the mechanisms responsible for each CNV from sequence data.

In an accompanying paper, high-resolution array comparative genomic hybridization was combined with whole genome sequencing data to find CNVs prevalent in Asians. DNA from 30 women—10 Korean, 10 Chinese, and 10 Japanese—was compared to three reference genomes. 20,099 CNVs were identified, an average of 670 CNVs per person. Notably, 2,183 of the 2,913 genes affected had not previously been identified as CNVs in a study of Europeans. Twenty-nine miRNA genes were included, and 35 potential gene fusions were identified.

The differences could be functionally significant. Copy number gains were seen in genes involved in nucleic acid metabolism and development, and copy number losses were seen in genes involved in cell adhesion. The authors conclude by insisting that “To more accurately apply CNV research to personalized medicine, copy number genotyping must not rely on relative copy number data, but should be able to identify the absolute copy number state in any given individual.” Something to aspire to.

View: Original Article