Genomic DNA extraction optimization and validation for genome sequencing using the marine gastropod Kellet’s whelk

Qiagen DNeasy blood and tissue kit

DNA extraction procedures followed the manufacturer’s protocol. For the Qiagen DNeasy Blood and Tissue kit, 25 mg tissue per sample was cut into small pieces and placed in a 1.5 mL microcentrifuge tube. 180 µL Buffer ATL and 20 µL Proteinase K were added. The sample was mixed by vortexing, then incubated at 56 °C overnight. The sample was vortexed for 15 s, then 200 µL Buffer AL was added. The sample was again vortexed, then 200 µL 100% ethanol was added. The sample was vortexed one last time, then pipetted into a DNeasy Mini spin column placed in a two mL collection tube. The sample was centrifuged at 6,000× g for 1 min and the flow-through was discarded. The DNeasy Mini spin column was placed in a new two mL collection tube, 500 µL Buffer AW1 was added, and then the sample was centrifuged at 6,000× g for 1 min. The flow-through was discarded, 500 µL Buffer AW2 was added, and then the sample was centrifuged at 20,000× g for 3 min. The flow-through and collection tube were discarded and the DNeasy Mini spin column was placed in a clean 1.5 mL microcentrifuge tube. 100 µL Buffer AE was added to the DNeasy membrane and incubated at room temperature for 1 min. The sample was centrifuged at 6,000 x g for 1 min to elute the DNA. The previous step was repeated with 100 µL Buffer AE to maximize DNA yield. Eluted DNA was stored at −20 °C. Protocol can be found in the SI 7.

Zymo Quick-DNA HMW Magbead kit

DNA extraction procedures followed the manufacturer’s protocol. For the Zymo Quick-DNA HMW Magbead kit, 25 mg tissue per sample was added to a 1.5 mL microcentrifuge tube. 95 µL DNA Elution Buffer, 95 µL Biofluid and Solid Tissue Buffer, and 10 µL Proteinase K were added to the sample. The sample mixture was pipette mixed 5 times and incubated at 55 °C for 1–3 h until tissue became soluble. The sample was centrifuged at 10,000× g for 1 min. The supernatant was removed and transferred to a new clean 1.5 mL microcentrifuge tube. 400 µL Quick-DNA MagBinding Buffer was added to the sample and mixed by pipetting. 33 µL MagBinding Beads were added to the sample, pipette mixed five times, and placed on a shaker for 10 min. The sample was placed on a magnetic stand until the beads separated from the solution. The supernatant was removed and discarded and the sample was taken off of the magnetic stand. 500 µL Quick-DNA MagBinding Buffer was added to the sample, pipette mixed five times, and placed on a shaker for 5 min. The sample was placed on a magnetic stand until the beads separated from the solution. The supernatant was removed and discarded and the sample was taken off of the magnetic stand. 500 µL DNA Pre-Wash Buffer was added to the sample and pipette mixed 10 times. The sample was placed on a magnetic stand until the beads separated from the solution. The supernatant was removed and discarded and the sample was taken off of the magnetic stand. 900 µL g-DNA Wash Buffer was added to the sample and pipette mixed 10 times. All liquid was transferred into a new microcentrifuge tube and placed on a magnetic stand until the beads separated from the solution.The supernatant was removed and discarded and the sample was taken off of the magnetic stand. The DNA wash step was repeated. The beads were dried for 20 min at room temperature and 50 µL DNA Elution Buffer was added. The sample mixture was pipette mixed 20 times and incubated at room temperature for 5 min. The sample was placed on a magnetic stand until the beads separated from the solution. The supernatant (eluted DNA) was removed and transferred to a new clean 1.5 mL microcentrifuge tube. Eluted DNA was stored at −20 °C. Protocol can be found in the SI 8.

PCI Protocol

The PCI DNA extraction procedure followed an adopted method from Thermo Fisher Scientific. For each sample, 30 mg tissue was homogenized mechanically in 1% SDS cell lysis buffer, 0.5M EDTA, and Proteinase K. The sample was then placed in a 65° C water bath for 1 h, and vortexed every 20 min. 1× volume Phenol Chloroform (PCI) was added to the sample and vortexed. The sample was centrifuged for 5 min at 16,000× g and the upper aqueous phase was transferred to a new tube. 20 µg Glycogen, 0.5× volume Ammonium Acetate 7.5 M, and 2.5× volume 100% ethanol were added to the sample. The sample was pipette mixed and kept at −20 °C overnight, then centrifuged at 4 °C for 30 min at 16,000× g and the supernatant was removed. 150 µL 70% ethanol was added to the sample, it was centrifuged at 4 °C for 2 min at 16,000× g and the supernatant was removed. The previous step was repeated. DNA was dried in a speedvac for 2 min and resuspended in 200 µL Elution Buffer. Eluted DNA was stored at −20 °C. Protocol can be found in the SI 9.

Salting out protocol

The Salting out protocol was adopted from Li et al. (2011). For each sample, 30 mg tissue was homogenized mechanically in 1% SDS cell lysis buffer, 0.5M EDTA, and Proteinase K. The sample was placed in a 65 °C water bath for 1 h, and vortexed for 5 s every 20 min. RNase A was added, the sample was lightly vortexed, then incubated at 37 °C for 10 min. 150 µL 7.5M Ammonium Acetate was added to the sample and mixed by vortexing. The sample was incubated at 4 °C for 5 min and centrifuged at 12,000× g for 10 min at 4 °C. The supernatant was transferred to a new tube and 60 µL 7.5M Ammonium Acetate was added. The sample was mixed by vortexing, incubated at 4 °C for 5 min, and centrifuged at 12,000× g for 10 min at 4 °C. The supernatant was transferred to a new tube and 1x volume 100% isopropanol was added. The sample was inverted 50 times and centrifuged at 8,000× g for 5 min at 4 °C. The supernatant was discarded and 400 µL 70% ethanol was added. The sample was inverted 3 times and centrifuged at 4 °C at 16,000× g for 5 min. The previous two steps were repeated. The sample was dried in a speedvac at 30 °C for 4 min. DNA was resuspended in 100 µL TE buffer and incubated at 4 °C overnight. Eluted DNA was stored at −20 °C. Protocol can be found in the SI 10.

Validation

The protocol selected as the most optimal among the four methods (based on the four selection criteria; see above and Results), was then validated on an independent set of 16 tissue samples from 9 adult and 7 recruit Kellet’s whelks collected in the wild at 15 sub-tidal reef locations across the species’ range in 2015, 2016, and 2017 (CDFW Scientific Collection Permit 8018 to C.W.) (Fig. 1). Tissue samples were collected from the foot of each individual Kellet’s whelk using a modified scalpel. Tissue samples from adults were collected non-lethally in the field, while recruits were collected whole in the field (because of their small size) and tissue samples were collected from them in the lab. Adult tissues and whole recruits collected in the field were frozen on dry ice or in liquid nitrogen, then transported to Cal Poly and stored at −80 °C until used for this study in 2021 and 2022. DNA extractions were evaluated based on DNA yield, DNA quality, and DNA purity. DNA yield and DNA quality were assessed using the same methods described above for comparing the four DNA extraction methods. DNA purity was measured by absorbance at 260/280 (contamination of proteins absorb at ∼280 nm, ideal ratio = ∼1.8) and 260/230 (contamination of salts, carbohydrates, and/or phenols absorb at ∼230 nm, ideal ratio = ∼1.8–2.2) (Thermo Fisher Scientific, 2009) using a Nanodrop-1000 spectrophotometer (Invitrogen, Waltham, MA, USA).

Scaling-up

The selected optimal method was applied to 3,325 Kellet’s whelk tissue samples and assessed by DNA yield and DNA quality using the same methods above. The tissue samples represent 2,026 adults and 1,299 recruits from the same set of wild-collected samples used in the validation step described above, but covering more sites (Table S1; Fig. 1). The extraction procedure was conducted by 13 undergraduate students at California Polytechnic State University in the Center for Applied Biotechnology from January 2022 to May 2022. DNA was extracted from samples in groups of 20 each week by each student. Three extractions were randomly selected from each set of 20 extractions bi-weekly for quality check assurance throughout the extraction procedure. Quality checks were conducted based on DNA yield and DNA quality, using the same methods described above for comparing the four DNA extraction methods. Sample quality and DNA concentration correlations were tested in R and visualized using ggplot2.

RAD-seq and GT-seq

A subset of the scaled-up DNA extractions were tested for sequencing quality on both GT-seq and RAD-seq platforms. A range of varying locations, years, and tissue types were used for each subset (Tables S3, S4, S5). For GT-seq, 96 DNA extractions were shipped frozen to Twin Falls, ID, USA and processed at the GTseek LLC laboratory. Briefly, exonuclease 1 was added to each gDNA extraction. gDNA was amplified using the set GT-seq primers. PCR products were indexed for Illumina sequencing with i7 and i5 index. Samples were normalized using Nate’s Plates Tagging and Normalization kit (SI 18). Size selection was conducted using SPRI beads. In depth methods can be found in the Supplementary Material section. Further optimization was conducted for GT-seq using the Zymo Clean and concentrator kit on aliquots of the same samples initially tested. The same methods were used for sequencing these samples. The genotyping pipeline used can be found at https://github.com/GTseq/GTseek_utils. For RAD-seq, 172 DNA extractions were shipped frozen to the Hawai‘i Institute of Marine Biology, Hawai‘i, USA and processed at the ToBo genomics lab (https://tobolab.org/). The gDNA digestion was performed with isoschizomer restriction enzymes (New England Bio-Labs, Ipswich, MA, USA). Libraries were generated using the KAPA Hyper Prep DNA kit (Roche Sequencing and Life Science) and then shipped frozen to the University of California, Davis, CA, USA (UCD) and sequenced at the UCD Genome Center through HiSeq2500. RAD-seq reads were filtered by their quality scores and the presence of the RADcutsite using process_radtags in Stacks (Rivera-Colón & Catchen, 2022). Protocols can be found in the SI 11–12.

Library preparation

To optimize and validate the selected DNA extraction method for next generation sequencing, a library from one of the DNA extractions generated using the Salting out protocol was prepared for Nanopore MinION sequencing using the rapid sequencing kit (SQK-RAD004; ONT, Oxford, UK) without any further cleaning or repair. Three DNA libraries from the same DNA extraction were prepared with further purification and concentration using the Genomic DNA Clean and Concentrator kit (Zymo Research, Irvine, CA, USA). The DNA was then selected for high molecular weight DNA using the PacBio Short Read Eliminator kit (PacBio, Menlo Park, CA, USA). Concentration was assessed using the dsDNA BR assay on a Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). Two libraries were prepared for Nanopore MinION sequencing using the Ligation Sequencing kit (SQK-LSK108; ONT, Oxford, UK) and NEBnext DNA Repair kit reagents (NEB, MA, USA) according to the manufacturer’s instructions. The other library was shipped frozen to the University of Oregon (Eugene, OR, USA) and prepared for Illumina Novaseq sequencing using the NEB Ultra II kit (NEB, Ipswich, MA, USA) according to the manufacturer’s instructions with a few modifications by the Genomics and Cell Characterization Facility (GC3F) at the University of Oregon (Eugene, OR, USA). DNA was mixed with the fragmentation reagents as described in the kit instructions and fragmented at 37 °C for 8 min, and end repaired as described in the manual. The end repaired sample was mixed with the ligation buffer and enhancer from the NEB kit, and 2.5 µL of 15 µm pre-annealed Tru-Seq style Y-adapter and ligated for 15 min at 20 °C. The sample was cleaned with 2, 0.75× bead cleans to remove adapter dimer. Samples were quantified by qPCR and loaded on the NovaSeq. Protocol methods can be found in the SI 13–17.

Whole genome sequencing

For Nanopore MinION sequencing, each library was loaded onto an R9 flow cell (FLO-MIN106, ONT). Priming and loading of the SpotON Flow Cells were performed using the standard protocol (sequencing gDNA (SQK-RAD004 or SQK-LSK109) Protocol) (Lu, Giordano & Ning, 2016). MinION sequencing was operated with MinKNOW v5.2.13 without basecalling (SI 17). Each flow cell was sequenced until <2 pores were sequencing. Basecalling was conducted on the raw Nanopore MinION reads using Guppy v6.2.1 and high accuracy mode (<5% error rate). Illumina Novaseq sequencing was conducted by GC3F at the University of Oregon using the NEBNext Ultra 2 DNA Library Prep kit for Illumina with slight modifications as stated above. All methods and protocols described above can be found in the Supplemental Information.

Extraction performance

Each extraction method was evaluated based on its difficulty, cost, yield, quality, and toxicity. Regarding DNA extraction difficulty (Table 1), the Qiagen kit was considered “easy” due to its lack of complex pipetting steps, while the Zymo kit and Salting out protocol were considered “medium” due to steps such as DNA washing and drying that required proficient lab skills. The PCI protocol was considered “hard” because it required steps with precise pipetting skills such as layered toxic solutions and under a fume hood. Comparing the cost per extraction of each method revealed the PCI and Salting out protocols, which required separate purchasing and assembly of reagents, to be approximately half as expensive as the Qiagen and Zymo kit methods, which contain pre-assembled reagents with the kit purchase (Table 1). DNA yield was broadly low in all samples prepared by the Qiagen Blood and Tissue kit (Table 1). DNA extractions produced by the Zymo kit produced optimal DNA yields (>1 µg of DNA) (Table 1). The Salting out protocol produced high and more variable DNA yields with a standard deviation of 9.45 µg (Table 1). The PCI method produced DNA yields lower than either the Zymo kit or salting out protocol, but higher than that by the Qiagen Blood and Tissue kit (Table 1).

The Qiagen and Zymo kits produced poor quality extractions. Samples prepared by the Qiagen Blood and Tissue kit had DNA degradation across most samples as observed by “digested” smears and low molecular weight bands on the agarose gel electrophoresis (Fig. 2). DNA extractions produced by the Zymo kit resulted in inconsistent HMW bands with DNA degradation (Fig. 2). The PCI method produced HMW bands with little DNA degradation (Fig. 2). The Salting out protocol produced consistent HMW bands without DNA degradation across all samples (Fig. 2). All methods had relatively low toxicity, except the PCI method which required the use of phenol and chloroform, which are dangerous chemicals that can be absorbed through the skin and are suspected mutagens. The Salting out protocol was selected as optimal among the four methods, due to its lack of difficulty, lowest cost per extraction, low toxicity, high DNA yield, and best quality HMW DNA.

Validation

To test the reliability of the Salting out protocol, it was validated on 16 Kellet’s whelk adult and recruit foot tissues from frozen samples collected across the species’ entire biogeographic range over multiple years (Table 2; Fig. 1). DNA yield of these extractions was broadly high, with an average of 35.7 ± 27.48 µg (Table 2). HMW bands were evident in all extractions, but signs of digestion/degradation were also evident (Fig. 3). The purity of these DNA extractions were evaluated using the samples absorbance ratios at A260/280 and A260/230. The average A260/280 ratio of these extractions was 1.85 ± 0.046 and A260/230 was 1.81 ± 0.27 indicating pure DNA with, potentially, some ∼230 nm absorbing contaminants.

Scaling

When the Salting out protocol was tested on 3,325 samples, the randomly sampled DNA extractions (252 DNA extractions from all different locations, tissue type, and years) resulted in a range of quality, with most falling into high enough quality for downstream GT-seq and RAD-seq applications. The overall quality of each extraction fell into three groups; Digested (8.7%), HMW + Degradation (29.8%), and HMW (61.5%) (Figs. 4; 5). The majority (81.3%) fell into the groups optimal for RAD-seq and GT-seq; high quality HMW group or HMW + Degradation. The remaining Digested samples (8.7%), would be the least likely to produce sufficient RAD-seq and GT-seq libraries. DNA concentration was found to be correlated with DNA quality (Fig. 4). This result is most likely due to smaller sample sizes that may have undergone DNA degradation more rapidly during the storing and thawing phase of tissue.

RAD-seq and GT-seq

A subset of 172 of the scaled-up DNA extractions were analyzed via RAD-seq, producing a total of 793,871,366 reads that passed quality assessment. Among the sequenced samples, nine (5.2%) exhibited a total read count lower than 410, with the remainder displaying a notably higher count of over 250,000 reads. All reads were further filtered by the presence of the anticipated RAD-seq cut site, resulting in an average of 78.9% of reads retained after all filtering from the raw reads (Fig. 6). A subset of 96 of the scaled-up DNA extractions were also analyzed with GT-seq. Twenty five (26%) generated >90% GT, 15 (15.6%) generated 30–90% GT, and 56 (58.3%) generated <30% GT (Fig. 6). Approximately half of the samples (49%) had raw read counts below 100,000 reads indicating unsuccessful sequencing. In general, the number of raw reads per sample decreased with lower % GT (Fig. 6). No correlation was found between successful genotyping and sample location or DNA yield (Table S6). After the same samples were cleaned using the Zymo Clean and Concentrator there was a 49% increase in successful genotype calling (>90%). Of the samples, 72 (75%) were above 90% GT, 13 (13.5%) were between 30 and 90% GT, and only 11 (11.5%) under 30% GT (Fig. 6). The number of raw reads across each sample was far more consistent with only one sample having less than 100,000 reads, a 50% increase in samples with optimal raw read counts after cleaning (Fig. 6).

Genome sequencing

The DNA from the Salting out protocol was sequenced on the Nanopore MinION and the Illumina Novaseq. The first library prepared using the rapid sequencing kit resulted in a failed sequencing run (Table 3). The sequencing run produced an estimated 14.01 k reads and 8.94 Mb bases, yet when finalized using the MinKONOW software, no data was produced (Table 3). Both Nanopore MinION libraries prepared using the Genomic DNA Clean Concentrator kit, the PacBio Short Read Eliminator kit, and the ligation sequencing protocol produced successful whole genome sequencing (Table 3). Both sequencing runs produced more than 80 times the estimated number of bases from the rapid sequencing kit run (Table 3). The KK_Ligationseq library produced a total of 2.6 Gb from 446.26 k reads and the KK_Ligationseq2 library produced a total of 1.4 Gb from 98.57 k reads of data that passed high quality base calling using Guppy. Each run had an estimated retention rate of 80.02% and 87.53% for reads that passed high quality filtering, respectively (Table 3).

The DNA from the salting out protocol extraction was also subjected to sequencing on the Illumina Novaseq. The same DNA cleaning and selection method used to produce successful libraries for nanopore sequencing was used. The Illumina Novaseq run produced 715.94 million reads and 113.83 Gb of bases. After filtering, there were 682.18 million reads and 104.85 Gb of bases. The filtered data consisted of 88.78% Q30 bases and 95.27% Q20 bases.