The fruits of a three-year effort have finally come to bear in the form of a new and improved sequencing method based on restriction-site associated DNA markers. A defining characteristic of a RAD tag is that one of its ends is fixed at a restriction site sequence while the other is sheared, allowing the surrounding genomic DNA to be sampled. Eric Johnson, an associate professor at the University of Oregon, originally developed the RAD marker method in 2005 when he applied it to microarrays; he has now expanded it to sequencing.
A recent paper published in PLoS One details how paired-end sequencing can be performed on RAD tags to extract a small number of reads from all the sequences generated, assemble those sequences into a longer contigs, and then repeat the process for each section of the genome. "This paper took a few years longer than we expected, but mostly because the Illumina sequencing platform kept getting better and we kept re-doing the proof-of-principle experiments to take advantage of the better read lengths and accuracy," Johnson says. "The variation for getting very long contigs of up to 5 kilobases involved combining the RAD paired-end sequencing with a long-insert library approach that involved circularization, and it took a few iterations to understand how to do that efficiently."
Johnson and his team used paired-end sequencing of RAD fragments to identify SNPs and determine the haplotype structure in the threespine stickleback, as well as sequence E. coli and stickleback genomic DNA with overlapping contigs of hundreds of nucleotides. The new method could prove to be a useful way of converting genome-wide short reads into individually assembled sequences stretching to thousands of nucleotides long.
Johnson says that in the near future, he would like to use the new RAD technique to expand cancer research and advance the development of treatments. "It is also possible to determine the haplotype of the assembled contig when the RAD sequence is heterozygous, so between these two aspects it is possible to get a very clear picture of the genome," he says. "Cancer genomes can be difficult to assemble if they have many structural variations and mutations, so RAD paired-end contigs could be used to help solve the structure of difficult genomes. That idea hasn't been tested yet, but that is something I would be interested in trying."