A DNA extraction protocol for improved DNA yield from individual mosquitoes

Introduction

DNA extraction for Anopheles mosquitoes is typically done using commercially available products (Brown et al., 2011; Demirci et al., 2012; Horton et al., 2010; Main et al., 2015; Norris et al., 2015; Weetman et al., 2012). They work well enough to provide a sufficient amount of DNA for PCR-based assays from a single mosquito. However, next-generation sequencing applications and high-throughput SNP genotyping assays exposed the limitation of DNA quantity from a single mosquito using typical extraction protocols (Marsden et al., 2011). Whole genome amplification could alleviate the issue but it also creates bias in genome representation (see Table 1 and results section below).

While trying to find alternative DNA extraction protocols for improved DNA yield but without compromising the automation option, we found that a combination of tissue lysis protocol from Life Technologies and DNA extraction protocol from Qiagen yielded higher DNA amount than the protocol using Qiagen or Life Technologies kit only.

Method

Results & Discussion

In our attempt to sequence whole genomes from field-collected individual mosquitoes, about 50% of specimens failed to pass the DNA quantity required (>30ng) for whole genome sequencing. These DNA samples were extracted using our established DNA extraction protocols using Qiagen kits and instruments (Lee et al., 2009; Main et al., 2015; Marsden et al., 2011; Norris et al., 2015; Slotman et al., 2006). The requirement of high genomic DNA content is not new. In the past, people have circumvented the problem by conducting whole genome amplification (Lee et al., 2013; Marsden et al., 2011; Weetman et al., 2012). We sequenced the whole genomes using original DNA in parallel with whole-genome amplified DNA to test if we can use whole genome amplification to bypass the DNA quantity issue.

Whole genome amplified DNA provided a higher number of reads than the original DNA with less DNA input (Table 1). However, comparison revealed that the particular whole genome amplification kit we used is not suitable for retrieving certain sections of genomes such as mitochondrial genome. This is indicated in the lower depth of coverage in mitochondrial genome in whole genome amplified material while the rest of chromosomes had higher depths relative to the library from original DNA. More importantly, the sequence generated from whole genome amplified samples produced number of inconsistent genotype calls (Table 2). This inconsistency became more apparent in mitochondrial sequences where heterozygous calls were produced where the genome sequence from the same original DNA had no such calls. These biases are likely introduced by the random primers used in the whole genome amplification kit. This result prompted us to pursue developing better DNA extraction protocols to improve DNA yield in an automated setup.

We found that the Life Technologies tissue lysis and extraction protocol (Table 3, line 4, in purple) was highly consistent in its DNA yield. A combination of the Life Technologies tissue lysis with the Qiagen BioSprint DNA extraction protocol (Table 3, line 3, in green) gave the highest average DNA yield (Wilcoxon rank sum test P-value=0.0031). The amount of magnetic beads added to tissue lysate had little effect on DNA yield. The amount of proteinase K (2µL vs 8µL) also showed little difference in DNA yield. Chemical lysis alone, without physical disruption, was not sufficient to produce consistency in DNA yield (Table 3, lines 5 and 6).

For a typical PCR-based assay, DNA quantity of 0.25–1.8 ng/µL in 200µL volume is sufficient. However, genomic approaches such as whole genome DNA library construction for next generation sequencing demand as little as 30ng of DNA. In our typical Qiagen BioSprint DNA extraction protocol, roughly 50% of DNA samples failed to yield 0.375ng/µL (Table 3, line 1, in blue), which leaves ~50µL of DNA for future study and allows for only a single trial of whole genome library construction. This constraint became a significant hindrance in our research involving whole genome sequencing from an individual mosquito.

This improved DNA extraction protocol will increase the chance of library construction from a single individual. Our observation is limited to trying out commercially available automated DNA extraction protocols and we do not have sufficient expertise on why certain protocols worked better or worse than others. As genomic approaches are more readily available to researchers, this improved DNA extraction protocol will facilitate such approaches that demand high-quantity DNA input from limited source material.

We have not rigorously tested all the possible combinations of extraction protocols. We only tested this on mosquito samples, and we only explored high-throughput automated DNA extraction protocols as we typically handle hundreds of mosquito samples at a time for population genetics studies. Therefore, our findings should be noted as suggestion for improving people’s own DNA extraction protocols and not as an advertisement of a commercially available product.

Data availability

F1000Research: Dataset 1. Raw data for ‘A DNA extraction protocol for improved DNA yield from individual mosquitoes’, 10.5256/f1000research.7413.d107517 (Nieman et al., 2015).