25 June 2009
Cost of genotyping plunges thanks to Sudoku logic
by Kate Melville
Sudoku, the hugely popular math-based game, may now be poised to revolutionize the fast-changing world of genome sequencing and medical genetics, claims a report in the journal Genome Research. Combining a 2,000-year-old Chinese math theorem with concepts from cryptology, Cold Spring Harbor Laboratory (CSHL) scientists have devised "DNA Sudoku," a methodology that allows tens of thousands of DNA samples to be combined and sequenced in parallel. Their achievement is in contrast to past approaches that, at best, can only combine hundreds of samples for sequencing.
"In theory, it is possible to use the Sudoku method to sequence more than a hundred thousand DNA samples," says CSHL Professor Gregory Hannon, leader of the team that invented the "Sudoku" approach. The new method promises to reduce costs dramatically. A sequencing project that costs upwards of $10 million using conventional methods may be accomplished for $50,000 to $80,000 using DNA Sudoku, he estimates.
Originally devised to overcome a sequencing limitation that dogged one of the Hannon lab's research projects, the new method has tremendous potential for clinical applications. It can be used, says Hannon, to analyze specific regions of the genomes of a large population and identify individuals who carry mutations that cause genetic diseases - a process known as genotyping.
The CSHL team has already begun to explore this possibility via a collaborative project with Dor Yeshorim, a New York-based organization that has collected DNA from thousands of members of orthodox Jewish communities. The organization's aim is to prevent genetic diseases such as cystic fibrosis that occur frequently within specific ethnic populations. The team's new method will now allow the many thousands of DNA samples gathered by Dor Yeshorim to be processed and sequenced in a single time-saving and cost-effective experiment, which should identify individuals who carry disease-causing mutations.
The mixing together and simultaneous sequencing of a massive number of DNA samples is known as multiplexing. In previous multiplexing approaches, scientists first tagged each sample with a barcode - a short string of DNA letters known as oligonucleotides - before mixing it with other samples that also had unique tags. After the sample mix had been sequenced, scientists could use the barcode tags on the resulting sequences as identification markers and thus tell which sequence belonged to which sample.
"But this approach is very limiting," explains researcher Yaniv Erlich. "It's time-consuming and costly to have to design a unique barcode for each sample prior to sequencing, especially if the number of samples runs in the thousands."
In order to circumvent this limitation, Erlich and others in the Hannon lab came up with the idea of mixing the samples in specific patterns, thereby creating pools of samples. And instead of tagging the individual samples within each pool, the scientists tagged each pool as a whole with one barcode. "Since we know which pool contains which samples, we can link a sequence to an individual sample with high confidence," says Erlich. The key to the team's innovation is the pooling strategy, which is based on the 2,000-year-old Chinese remainder theorem.
The method, which the CSHL team has patented, is currently best suited for genotype analyses that require only short segments of an individual's genome to be sequenced to find out if the individual is carrying a certain variant of a gene or a rare mutation. But as sequencing technologies improve and researchers gain the ability to generate sequences for longer segments of the genome, Hannon envisions wider clinical applications for their method.
Source: Cold Spring Harbor Laboratory