Protocol: a versatile, inexpensive, high-throughput plant genomic DNA extraction method suitable for genotyping-by-sequencing

Abstract

Background: The recent development of next-generation sequencing DNA marker technologies, such as genotyping-by-sequencing (GBS), generates thousands of informative single nucleotide polymorphism markers in almost any species, regardless of genomic resources. This enables poorly resourced or "orphan" crops/species access to high-density, high-throughput marker platforms which have revolutionised population genetics studies and plant breeding. DNA quality underpins success of GBS methods as the DNA must be amenable to restriction enzyme digestion and sequencing. A barrier to implementing GBS technologies is access to inexpensive, high-throughput extraction methods that yield sequencing-quality genomic DNA (gDNA) from plants. Several high-throughput DNA extraction methods are available, but typically provide low yield or poor quality gDNA, or are costly (US$6-$9/sample) for consumables.

Results: We modified a non-organic solvent protocol to extract microgram quantities (1-13 μg) of sequencing-quality high molecular weight gDNA inexpensively in 96-well plates from either fresh, freeze-dried or silica gel-dried plant tissue. The protocol was effective for several easy and difficult-to-extract forage, crop, horticultural and common model species including Trifolium, Medicago, Lolium, Secale, Festuca, Malus, Oryza, and Arabidopsis. The extracted DNA was of high molecular weight and digested readily with restriction enzymes. Contrasting with other extraction protocols we assessed, Illumina-based sequencing of GBS libraries developed from this gDNA had very uniform high quality base-calls to the end of sequence reads. Furthermore, DNA extracted using this method has been sequenced successfully with the PacBio long-read platform. The protocol is scalable, readily automated without requirement for fume hoods, requires approximately three hours to process 192 samples (384-576 samples/day), and is inexpensive at US$0.62/sample for consumables.

Conclusions: This versatile, scalable and simple protocol yields high molecular weight genomic DNA suitable for restriction enzyme digestion and next-generation sequencing applications including GBS and long-read sequencing platforms such as PacBio. The low cost, high-throughput, and extraction of high quality gDNA from a range of fresh and dried source plant material makes this method suitable for many sequencing and genotyping applications including large-scale sample screening underpinning breeding programmes.

Keywords: Arabidopsis; Festuca; Freeze-dried; High-throughput; Lolium; Malus; Medicago; Next-generation sequencing; Oryza; Secale; Silica gel-dried; Trifolium.

Fig. 1

Genomic DNA (gDNA) extracted and restriction endonuclease-digested from forage legumes and grasses, and horticultural species. Samples of gDNA a extracted from freeze-dried leaf tissue or grass pseudostem using the high-throughput protocol and b digested with HindIII restriction endonuclease were resolved and visualised by electrophoresis in an agarose lithium borate buffer (0.8% w/v) gel containing 25 μg ethidium bromide. The samples were, in order: (1) 1 kb Plus size ladder (www.thermofisher.com); (2) Trifolium repens (white clover); (3) T. occidentale (western clover); (4) T. pallescens (pale clover); (5) T. pratense (red clover); (6) T. subterraneum (subterranean clover); (7) T. uniflorum (8) Medicago sativa (alfalfa/lucerne); (9) Lolium perenne (perennial ryegrass); (10) Festuca arundinacea (tall fescue); (11) Secale cereale (rye); (12) Malus pumila (apple); (13) λ DNA standard, and (14) 1 kb Plus size ladder. The λ DNA standard was digested with HindIII in (B)

Fig. 2

Genomic DNA (gDNA) extracted from fresh, freeze-dried, and silica gel-dried plant tissue. Samples of gDNA extracted from a clover leaf tissue or grass pseudostem b rice pseudostem and c Arabidopsis thaliana leaf tissue using the high-throughput protocol were resolved and visualised by electrophoresis in an agarose lithium borate buffer (0.8% w/v) gel containing 25 μg ethidium bromide. F fresh, Fd freeze-dried, Sd silica gel-dried; Tr Trifolium repens (white clover); Tp T. pratense (red clover); Lp Lolium perenne (perennial ryegrass); Fa Festuca arundinacea (tall fescue); Os Oryza sativa (rice); At Arabidopsis thaliana; λ λ DNA standard. The samples were flanked by 1 Kb Plus size ladders (www.thermofisher.com)

Fig. 3

A typical DNA extraction from freeze-dried and fresh tissue using the 96-well plate method. Genomic DNA was extracted from 96 individuals of a Trifolium repens (white clover; freeze-dried leaf) and b Lolium perenne (perennial ryegrass; fresh pseudostem) using the 96-well plate protocols, and resolved and visualised by electrophoresis on an agarose lithium borate buffer (0.8% w/v) gel containing 25 μg ethidium bromide. The samples were flanked at either end by 1 Kb Plus size ladders (www.thermofisher.com). One T. repens individual had a poor DNA yield (top row, ninth lane from the right) and was subsequently re-extracted using the Individual Tube freeze-dried tissue protocol. These DNA samples have been used for developing genotyping-by-sequencing libraries and yielded high-quality sequence data

Fig. 4

Influence of genomic DNA (gDNA) extraction protocols on sequencing quality. We assessed and modified methodologies to establish a high-throughput protocol to extract gDNA from white clover, a species prone to yielding degraded DNA. Chemistry based on a CTAB protocol [6] was modified for gDNA extraction using 96-well plates (a). Adaptation of the Whitlock method [2] to a 96-well plate protocol (b), and development of a streamlined inexpensive protocol described in this paper (c). In these examples, gDNA was extracted from freeze-dried leaves and aliquots (2 μL from 100 μL gDNA extraction/elution) were resolved and visualised by electrophoresis in an agarose lithium borate buffer (0.8% w/v) gel containing 25 μg ethidium bromide. The samples were flanked at either end by 1 kb Plus size ladders (www.thermofisher.com). Sequence quality of the extracted gDNA was assessed by producing a genotyping-by-sequencing (GBS) library [1] comprising 95 individuals from each of the white clover gDNA extractions shown above (a–c). The GBS libraries were single-end sequenced (100 bp) on a single lane each of an Illumina 2500 Hi-Seq sequencer. Sequencing quality assessment using FastQC version 0.10.1 [9] for GBS libraries made from the gDNA shown a–c is represented in graphs describing quality across all bases from every sequence read at each position (d–f, respectively). Sequence quality is based on phred scores [10], an exponential scale where, for example, 20 = one incorrect sequence base-call in 100, and 30 = one incorrect base-call in 1000. The y-axis shows the quality scores, and the higher the score the greater confidence in the base-calls at that position. The background of the graph divides the y-axis into very good quality calls (green), reasonable quality (orange), and poor quality (red)