Regular readers of the Sage Science blog know that we can never resist a good methods paper. We enjoyed this publication from the phyloinformatics group at RIKEN detailing a modified protocol for generating high-quality mate-pair libraries while significantly reducing costs.
Lead author Kaori Tatsumi and colleagues focused on Illumina’s Nextera Mate Pair Sample Prep Kit, a workhorse for this application. They rely on mate-pair sequencing to improve the contiguity of de novo assemblies, particularly for non-model organisms. The Nextera kit “has significantly reduced the difficulty, preparation time and cost of preparing mate-pair libraries,” the scientists report, but “there remain opportunities to improve the efficiency and reduce costs for this preparation technique.”
The team tried several tactics to achieve this goal. They started with reducing the amount of enzyme used in the workflow, and also concocted their own homemade buffer. Tests showed that these worked at least as well as the original protocol. “The use of a reduced volume of enzyme and self-made buffer allows for a higher number of tagmentation reactions to be attempted using variable conditions,” the scientists write. “In particular, this modification, which yields larger DNA molecules, is advantageous in preparing mate-pair libraries spanning large distances (>10 kb), which tends to be hindered by low yields.”
They also swapped the manufacturer’s recommended order of the strand displacement step and the size selection step, sizing DNA first on the BluePippin and then performing strand displacement. “Reversing the order of strand displacement and size selection enables significantly smaller volumes of strand displacement polymerase and buffer for reduced amounts of size-selected DNA, which allows for a larger number of preparations (up to 3-fold) than the standard protocol,” Tatsumi et al. found. The revised process still yielded enough library volume for the rest of the pipeline.
Finally, they added more shearing steps to improve accuracy of reads recognized as mates, and adjusted the number of cycles of sequencing on an Illumina HiSeq. Compared with libraries prepared according to the original protocol, their method resulted in much longer scaffold lengths and a higher percentage of genes covered completely in the assembly. The revised protocol also costs significantly less than standard preparation.
Kudos to the RIKEN team for their hard work on a cool new method!