Introduction

Multiple displacement amplification (MDA) is a method of amplifying DNA from small amounts of template DNA: double-stranded DNA is denatured (e.g. via increased pH) before the addition of random hexamer primers and a strand-displacing isothermal polymerase (e.g. phi29 DNA polymerase). During the amplification reaction, 5’ ends of synthesised strands become displaced by the growing 3’ end of other nascent strands. The displaced strands themselves then act as templates onto which new hexamers anneal and extend (Figure 1). This process continues and the template DNA is amplified. The method can generate many micrograms of product from a few picograms of template. However, libraries generated from such products typically perform poorly in sequencing, resulting in low flow cell output and reduced quality score. In this document, we explain how we improve data output and quality using an optimised protocol for the Ligation Sequencing Kit:

Ligation sequencing gDNA V14 - whole genome amplification (SQK-LSK114)

When sequencing MDA products on Oxford Nanopore devices, we observe reduced flow cell output and populations of reads with low quality score (Qscore), typically reducing median Qscore by ~10%. We expect this to be caused by the highly branched structure of the MDA product. To resolve this, we treated MDA products with T7 endonuclease I, which cuts at single-stranded regions (Figure 1). We found that this helps increases flow cell output by reducing the frequency of terminal blocking, without causing a significant reduction in observed read length (Figure 2) and helps restore the quality score to the levels observed with native genomic DNA.

Figure 1. Overview of the MDA method with T7 endonuclease I.

Figure 2. Impact of T7 endonuclease I treatment on MDA products. a) Libraries were sequenced on MinION Flow Cells and output was improved by T7 endonuclease I treatment. b) Read length shows a slight reduction is observed after T7 endonuclease I treatment.

Sequencing of ZymoBIOMICS Microbial Community DNA Standard

We evaluated the performance of our MDA protocol using the ZymoBIOMICS Microbial Community DNA Standard. This standard includes a range of bacterial genomes to provide a robust benchmark for performance assessment, and is particularly useful for detecting bias.

We titrated the input mass of the ZymoBIOMICS Microbial Community DNA Standard to evaluate the performance of the MDA protocol under conditions of low template input. Additionally, we compared the Ligation Sequencing Kit, the Rapid Sequencing Kit and the Rapid PCR Barcoding Kit 24 to determine the best library preparation method for sequencing of MDA products. Libraries were sequenced on MinION Flow Cells.

Results

Output and read length

Libraries generated using the Ligation Sequencing Kit with the optimised protocol for products of MDA generated slightly less sequencing output than libraries generated using the native template (Figure 3a). We also observed that libraries prepared from the products of MDA using the Rapid Sequencing Kit generated significantly less output than the Ligation Sequencing Kit approach. Therefore, we do not recommend using the Rapid Sequencing Kit in preparing libraries after MDA. Template input into the MDA reaction is also reduced and the libraries generated using the Ligation Sequencing Kit produced a decreased flow cell output (Figure 3b). We speculate that with reduced template inputs, an increase in random priming events occur, leading to complex structures which can block pores and reduce flow cell output. Flow cell output can be recovered from low input samples by performing an additional PCR amplification using the Rapid PCR Barcoding Kit 24 after the MDA step (Figure 3c). We postulate that the complex structures generated at low input by MDA do not proliferate in the PCR, and the resulting library consists of molecules less likely to cause blocking. Read length distributions for the recommended approaches are shown in Figure 4.

Figure 3. Output of libraries generated using the ZymoBIOMICS Microbial Community DNA Standard with and without MDA. a) Libraries generated using the Ligation Sequencing Kit with MDA products (MDA-LSK) generated slightly less sequencing output than the native template. We also observed that libraries prepared from the products of MDA using the Rapid Sequencing Kit (MDA-RAD) generated less output than the Ligation Sequencing Kit (MDA-LSK). b) As the template input into the MDA reaction is reduced, the libraries generated produced a decreased flow cell output. c) Performing an additional PCR (MDA-RPB) after MDA can increase flow cell output, especially at low template input.

Figure 4. Read length of libraries generated using the ZymoBIOMICS Microbial Community DNA Standard with and without MDA. The read length distribution for MDA libraries (prepared with either LSK or RPB) varies little when template input is reduced (not shown).

Coverage

We mapped the sequencing data to the ZymoBIOMICS Microbial Community DNA Standard reference sequences and compared the alignment coverage to each genome to the theoretical composition (Figure 5 and Table 1. Note: a suitable reference for Crytococcus neoformans is not available and therefore has been excluded from the subsequent analysis). We observed that the composition of sequencing data from the native DNA (not amplified) closely matches the theoretical composition of the community reference. Libraries generated using MDA product resulted in significant bias, leading to an increase in abundance of some bacterial strains, and a significant reduction in others. Given that sequencing of the native DNA shows a good correlation between expected and observed community composition, we assume that the bias is caused by the MDA rather than the library preparation or sequencing. We did not detect an increase in bias due to reduced template input into the MDA reaction, nor did we detect a significant additional bias introduced following the PCR amplification of MDA products.

Figure 5. Expected and observed composition of constituents of the ZymoBIOMICS Microbial Community DNA Standard with and without MDA.

Table 1. Expected and observed composition of constituents of the ZymoBIOMICS Microbial Community DNA Standard with and without MDA.

Summary

When sequencing the products of MDA, we recommend the use of the Ligation Sequencing Kit, rather than the use of the Rapid Sequencing Kit to prepare libraries, to maximise flow cell output. T7 endonuclease I treatment of MDA products is key to improving the performance of the downstream sequencing library. In instances where low flow cell output is observed when sequencing products of MDA, an additional PCR amplification using the Rapid PCR Barcoding Kit 24 can help improve output. Libraries prepared from products of MDA display some coverage bias; minimal changes in bias were observed with variations in template inputs for the MDA reaction or with subsequent PCR amplification of MDA products.