A new look at the human genome suggests that unappreciated variations in its fundamental architecture, rather than point-by-point mutations, may be responsible for most genetic difference among people.

Point-by-point mutations, called single nucleotide polymorphisms, involve simple changes to DNA lettering. They’re the best-studied type of variation, the target of most genomic disease hunts, and the substance of commercially available personal genome readouts.

More complex yet less-studied are structural variations, which involve large-scale changes: wholesale duplications and reversals, or unexpected additions and omissions, of long DNA sequences.

Traditional genome sequencing techniques are too fuzzy and piecemeal to make sense of these, yet “our observations suggest that structural variations are more specific to individuals than single nucleotide polymorphisms are,” wrote researchers led by Jun Wang of the Beijing Genomics Institute in a July 24 Nature Biotechnology study.

It might seem counterintuitive that big changes have been harder to detect than small ones, but it’s a consequence of how genomes are read. Every method involves breaking long DNA sequences — the human genome contains three billion DNA pairs — into pieces, then trying to reassemble them. The methods vary according to fragment size and reassembly technique, but as a rule it’s far less expensive and time-intensive to use small fragments.

As a result, most genomic studies, including gold-standard genome-wide association surveys, involve sequences reassembled from small pieces. As with a jigsaw puzzle or a book, however, larger fragments would work better. If the pieces are too small, or the text blocks just a few letters long, it’s difficult to be certain what the final product ought to look like. It’s possible to compare two pieces, but not puzzle sections or paragraphs.

“One reason you’ve heard more about single nucleotide polymorphisms, that they’ve come to the fore even though they’re a more minor form of variation than these structural variants, is that they were easier to see,” said Yale University bioinformaticist Mark Gerstein, who was not involved in the study.

In the new study, Wang and colleagues used algorithms that assemble long, relatively intact genome sequences from small fragments, allowing them to see more structural variation than is usually possible. In a high-profile earlier study, they’d used it to sequence a giant panda genome; this time they compared structural variations across 106 people from the 1000 Genomes Project.

They found that individuals seem to be distinguished less by their SNPs than their structural variations. “Defining structural variations will be of considerable importance for future analyses of personal genomes,” they wrote.

The new study didn’t attempt to link structural variation with traits or disease, but it’s possible that as-yet-unstudied structural variations contain a portion of what geneticsts call the “missing heritability.” Of common disease risk that’s clearly hereditary, only a small fraction can currently be traced to genes.

“Structural variations are a very hot topic,” said Gerstein. “They’re probably the major form of variation in humans.”

Image: A comparison of structural variation in an Asian genome (left) and one African genome (right). (Wang et al/Nature Biotechnology)

See Also:

Citation: “Structural variation in two human genomes mapped at single nucleotide–resolution by whole genome de novo assembly.” By Yingrui Li, Hancheng Zheng, Ruibang Luo–, Honglong Wu, Hongmei Zhu, Ruiqiang Li, Hongzhi Cao, Boxin Wu, Shujia Huang, Haojing Shao, Hanzhou Ma, Fan Zhang, Shuijian Feng, Wei Zhang, Hongli Du, Geng Tian, Jingxiang Li, Xiuqing Zhang, Songgang Li, Lars Bolund, Karsten Kristiansen, Adam J de Smith, Alexandra I F Blakemore, Lachlan J M Coin, Huanming Yang, Jian Wang & Jun Wang. Nature Biotechnology, July 24, 2011.