Improved workflows for high throughput library preparation using the transposome-based Nextera system

Abstract

Background: The Nextera protocol, which utilises a transposome based approach to create libraries for Illumina sequencing, requires pure DNA template, an accurate assessment of input concentration and a column clean-up that limits its applicability for high-throughput sample preparation. We addressed the identified limitations to develop a robust workflow that supports both rapid and high-throughput projects also reducing reagent costs.

Results: We show that an initial bead-based normalisation step can remove the need for quantification and improves sample purity. A 75% cost reduction was achieved with a low-volume modified protocol which was tested over genomes with different GC content to demonstrate its robustness. Finally we developed a custom set of index tags and primers which increase the number of samples that can simultaneously be sequenced on a single lane of an Illumina instrument.

Conclusions: We addressed the bottlenecks of Nextera library construction to produce a modified protocol which harnesses the full power of the Nextera kit and allows the reproducible construction of libraries on a high-throughput scale reducing the associated cost of the kit.

Figure 1

Post-normalisation DNA concentrations. DNA samples with input concentrations of approximately 20 ng/μl (×) and 10 ng/μl (□) were normalised with the Axygen PCR Normaliser kit. The output concentration is shown for each genome across a range of GC contents.

Figure 2

Library QC Tape station electropherogram. Nextera Post-PCR libraries constructed with a range of concentrations (post-normalisation) and gDNA from four different genomes: Mycobacterium tuberculosis (purple), Escherichia coli (blue), Clostridium difficile (red) and Plasmodium falciparum (green). Libraries were constructed using the standard Nextera protocol (A). Evaluation of the Axyprep Mag Normaliser kit (B): two individual C. difficile Nextera libraries were constructed using the standard Illumina protocol (light/dark green) and two with our normalisation workflow (light/dark purple). Where the standard library had very short inserts, our method produced a library with the normal size distribution. Evaluation of C. difficile Nextera Post-PCR libraries constructed using varying volume Nextera reactions (C): standard (purple), half-volume (green), quarter-volume (red) and one-eighth volume (blue). Size distribution profiles of libraries constructed using normalisation followed by reaction E (D): M. tuberculosis (purple), E. coli (blue), C. difficile (red) and P. falciparum (green).

Figure 3

Library insert size. Libraries were constructed using the standard (A) or reaction E Nextera (B). Sequencing metrics showed the library insert sizes for C. difficile (i), E. coli (ii) and M. tuberculosis (iii) to be approximately 250-300 bp irrespective of the original TapeStation profile (Figure 2A and B). C. difficile produced a library with a shorter size distribution (~180 bp) in this experiment (Bi).

Figure 4

Sequencing QC: genome coverage. Coverage across the genome for C. difficile(A,B), E. coli(C,D) and M. tuberculosis(E,F) libraries constructed with standard (A,C,E) and reaction E (B,D,F) Nextera preparation. The coverage is consistent between the two preparations, with both protocols producing good, even coverage across the genome. The large spikes seen in all three genomes probably represent repetitive regions and are consistent between the two preparations.