ThruPLEX FLEX outperforms NEBNext Ultra II for DNA-seq of cell-free and FFPE DNA
Note: The protocols and QC procedures for ThruPLEX HV kits have been updated to accommodate lower inputs and compatibility with the Unique Dual Index Kit sets. While product naming has been revised accordingly (ThruPLEX FLEX), reagent formulations remain unchanged.
A next-generation sequencing (NGS) library preparation system must provide a simple, streamlined workflow that accommodates a wide range of sample inputs without compromising accuracy. ThruPLEX DNA-Seq FLEX satisfies these requirements with a comprehensive and reliable method that enables the generation of high-complexity libraries. The kit's unique single-tube workflow makes this one of the fastest and most consistent library preparation systems on the market. This single-tube approach also helps prevent sample loss by eliminating the need for the time-consuming intermediate bead purification steps required by many competitor kits (Figure 1). Additionally, ThruPLEX DNA-Seq FLEX saves time by eliminating the need for adapter dilution and protocol optimization (Figure 1).
By increasing the starting input volume (30 µl) and range of input (up to 200 ng) relative to the original ThruPLEX DNA-Seq Kit, we have allowed for the production of high-complexity libraries without the need for an upfront sample concentration step. Libraries generated with ThruPLEX DNA-Seq FLEX can be used directly for whole-genome sequencing or enriched using any of the leading platforms for targeted sequencing applications.
ThruPLEX DNA-Seq FLEX
Kapa Hyper Prep
NEBNext Ultra II
Hands-on time
15 min
20 min
20 min
Total time
2.4–2.6 hr
2.5–2.7 hr
3.1–3.2 hr
Single-tube workflow
Yes
No
No
Adapter dilution
No
Yes
Yes
Intermediate cleanup
No
Yes
Yes
Post-ligation size selection
No
No
Yes (>100 ng)
Figure 1. Comparison of three leading NGS library preparation chemistries. Panel A. Total time is representative of the time required to amplify inputs of 200 and 5 ng with each chemistry to yield enough Illumina-compatible dual-indexed libraries for target enrichment. ThruPLEX DNA-Seq FLEX is the only single-tube workflow and the only chemistry which does not require adapter dilution, intermediate cleanup, or post-ligation size selection for both challenging and more standard sample types. The single-tube worklow provides the quickest and simplest path for obtaining sequencing-ready libraries with the least amount of hands-on time. Panel B. The ThruPLEX FLEX workflow consists of three simple steps that take place in the same well or PCR tube, eliminating the need to purify or transfer the sample material.
Results
Competitive library coverage uniformity
The ThruPLEX DNA-Seq FLEX workflow consists of three simple steps that take place in a single well or PCR tube—with just 15 minutes of hands-on time—to yield indexed libraries from fragmented DNA within two hours. Increasing the input volume (30 µl) and range of input (up to 200 ng) at the start of the protocol enables the generation of higher-complexity libraries for sequencing or target enrichment. Generation of high-complexity libraries is critical for achieving even coverage across the genome for whole-genome sequencing (Figure 2). When compared to NEBNext Ultra II, libraries generated using ThruPLEX DNA-Seq FLEX provide coverage closer to the ideal normalized coverage.
Input
Total reads
Reads aligned
Duplicate
ThruPLEX FLEX (formerly ThruPLEX HV)
50 ng
7,868,884
97%
0.73%
5 ng
7,796,764
97%
0.84%
NEBNext Ultra II
50 ng
7,922,699
97%
0.98%
5 ng
7,978,001
97%
0.89%
Figure 2. Superior coverage uniformity.Panels A and B. Libraries were prepared in triplicate from 5-ng and 50-ng inputs of a quantitative multiplex reference standard consisting of gDNA pooled from HCT116, RKO, and SW48 cell lines (Horizon Discovery). Libraries were generated following the ThruPLEX DNA-Seq FLEX (formerly ThruPLEX HV) or NEBNext Ultra II (NEBNext) protocols. Paired-end sequencing was performed on a NextSeq® 500/550 Mid Output Kit v2.5 (150 Cycles), and data were downsampled to 8 million total reads. The vertical gray bars represent the expected GC content distribution using 100-bp windows.
Target enrichment with FFPE and formalin-compromised inputs
Processing of clinical research materials for long-term storage may include fixation with formalin. Exposing samples to formalin can lead to significant damage of the nucleic-acid content, which is often present in limited quantities. Due to this damage, the construction of libraries can prove challenging and, therefore, generally requires a kit with a robust repair mechanism to produce enough post-PCR product for target enrichment. ThruPLEX DNA-Seq FLEX was designed to accommodate a large input volume and a higher amount of starting material, which improves coverage and mutation detection by increasing the complexity of the input. Another important consideration for the confident calling of low-frequency mutations is achieving even coverage throughout the genome in order to ensure optimal read depth at all relevant loci. To facilitate even coverage, our system has been optimized for improved coverage uniformity across a broad range of inputs with varying levels of damage and GC content.
We compared the performance of ThruPLEX DNA-seq FLEX and NEBNext Ultra II (Figure 3) on libraries generated in triplicate with 50-ng and 5-ng inputs of Horizon DNA references, including formalin-compromised material with severe damage, as well as with 30-ng and 5-ng of formalin-fixed paraffin-embedded material. ThruPLEX FLEX libraries outperformed NEBNext libraries in mean target coverage for the formalin-compromised DNA and exhibited comparable mean target coverage for the FFPE samples (Figures 4). Both kits performed similarly in the detection of positive variants, on the two types of formalin-treated samples (Figure 5).
Input type
Chemistry
Input amount
Reads
Reads aligned
Bases aligned
On-target bases
Formalin compromised (severe)
ThruPLEX DNA-Seq FLEX
50 ng
4.93 x 106
98.0%
3.6 x 108
1.8 x 108
5 ng
4.86 x 106
97.6%
3.4 x 108
1.3 x 108
NEBNext Ultra II
50 ng
5.00 x 106
98.7%
3.1 x 108
1.6 x 108
5 ng
5.00 x 106
98.6%
3.1 x 108
1.2 x 108
Formalin-fixed paraffin-embedded
ThruPLEX DNA-Seq FLEX
30 ng
4.85 x 106
97.6%
3.0 x 108
8.7 x 108
5 ng
4.66 x 106
97.0%
2.9 x 108
8.2 x 108
NEBNext Ultra II
30 ng
5.00 x 106
98.3%
3.1 x 108
8.9 x 108
5 ng
5.00 x 106
98.3%
3.1 x 108
9.0 x 108
Figure 3. Excellent target-capture efficiency across a range of DNA quality. Libraries were generated in triplicate with 50-ng and 5-ng inputs of Horizon DNA references using formalin-compromised material with moderate and severe damage (Cat. # HD803), as well as with 30-ng and 5-ng inputs of formalin-fixed paraffin-embedded material (Cat. # HD200). Libraries were amplified with ThruPLEX DNA-Seq FLEX chemistry or NEBNext Ultra II. Furthermore, on-target efficiency was roughly equivalent between the two kits (data not shown).
Figure 4. ThruPLEX FLEX has better mean target coverage for formalin-compromised and FFPE-treated references than NEBNext Ultra II. Libraries were constructed in triplicate with ThruPLEX DNA-Seq FLEX (formerly ThruPLEX HV, indicated in blue) and NEBNext Ultra II using formalin-compromised material with moderate and severe damage, as well as with formalin-fixed paraffin-embedded material. Amplified libraries were purified with AMPure beads and pooled for target capture with IDT's xGEN Pan Cancer Panel covering 800 kb of the human genome. Paired-end sequencing was performed on a NextSeq 500/550 Mid Output Kit v2.5 (150 Cycles). Data were downsampled to 5 million total reads per sample and aligned to hg19 using bowtie2. Variants were called using VarDict Variant Caller and sorted to include a minimum of 30X coverage and allele frequencies above 0.5%.
Figure 5. ThruPLEX FLEX detects more mutations in FFPE samples than NEBNext Ultra II. Libraries were constructed in triplicate with ThruPLEX DNA-Seq FLEX (formerly ThruPLEX HV) and NEBNext Ultra II using formalin-compromised material with moderate and severe damage, as well as with formalin-fixed paraffin-embedded material. Targets were enriched using the xGEN Pan Cancer Panel (IDT). A positive detection occurs when at least two of three replicates positively detect a variant.
Target enrichment with cell-free DNA
Sequencing of cell-free DNA (cfDNA) faces similar challenges to that of formalin-compromised and FFPE samples. We, therefore, set out to test the performance of ThruPLEX DNA-Seq FLEX with wild-type cfDNA and cfDNA containing eight confirmed single-nucleotide variants (SNVs) occurring at 5% allelic frequency. While mean target coverage was comparable between ThruPLEX FLEX and NEBNext Ultra II libraries (Figure 6), ThruPLEX FLEX enabled detection of more positive variants in both the wild-type cfDNA and cfDNA reference containing confirmed SNVs (Figure 7).
Figure 6. Mean target coverage is comparable for ThruPLEX FLEX and NEBNext Ultra II libraries. Libraries were constructed in triplicate with ThruPLEX DNA-Seq FLEX (formerly ThruPLEX HV) and NEBNext Ultra II using 50 and 5 ng of a wild-type multiplex cfDNA reference standard (Horizon Discovery; Cat. # HD776) and a reference standard with eight additional mutations occurring at a 5% allelic frequency. Targets were enriched using the xGEN Pan Cancer Panel (IDT).
Figure 7. ThruPLEX FLEX detects more positive variants in wild-type cfDNA and reference cfDNA with SNVs confirmed 5% allele frequencies. Libraries were constructed in triplicate with ThruPLEX DNA-Seq FLEX (formerly ThruPLEX HV) and NEBNext Ultra II using 50-ng and 5-ng inputs of a wild-type multiplex cfDNA reference standard and a reference standard with eight additional mutations at a 5% allelic frequency. Targets were enriched using the xGEN Pan Cancer Panel (IDT). A positive detection occurs when at least two of three replicates positively detect a variant.
Conclusion
ThruPLEX DNA-Seq FLEX is a simple, fast, and accurate system which employs a three-step protocol that can be completed in a single tube in two hours. The ThruPLEX DNA-Seq FLEX library preparation kit for Illumina sequencing elevates the ThruPLEX product family by accommodating a larger input volume and greater amount of starting material than previous ThruPLEX DNA-seq kits. Along with these improvements, ThruPLEX DNA-Seq FLEX retains the coveted single-tube ThruPLEX workflow with no intermediate cleanup steps. Through workflow optimization and reagent reformulation, ThruPLEX DNA-Seq FLEX outperforms NEBNext Ultra II in coverage of regions with increasing GC content as well as in the detection of variants in both FFPE and cell-free DNA samples.
Methods
DNA preparation
Human genomic DNA from Horizon Discovery (Cat. #s HD701, HD803, and HD200) was sheared on a Covaris M220 following the 250-bp shearing protocol. Sheared input material and cfDNA material not requiring shearing (Horizon Discovery, Cat. #s HD776 and HD777) were evaluated for correct size on an Agilent 2100 BioAnalyzer using Agilent High Sensitivity DNA Reagents. The concentration of these samples was measured using a Qubit 2.0 Fluorometer with the Quant-iT dsDNA Assay kit, high sensitivity (Thermo Fisher Scientific).
Library preparation
Libraries were prepared following the manufacturer's instructions using the ThruPLEX DNA-Seq FLEX kit or NEBNext Ultra II kit. All libraries were generated using dual indexes. Amplified libraries were purified using AMPure XP beads (Beckman Coulter) and eluted in low-TE buffer for whole-genome sequencing (WGS). Purified library size was assessed on the Agilent 2100 BioAnalyzer using Agilent High Sensitivity DNA Reagents. Libraries were quantified by qPCR using the Library Quantification Kit (Takara Bio, Cat. # 638324) or Qubit 2.0 Fluorometer with the Quant-iT dsDNA Assay kit, high sensitivity (Thermo Fisher Scientific).
Target capture
Amplified libraries were purified with AMPure beads and pooled for target capture with the IDT xGEN Pan Cancer Panel covering 800 kb of the human genome.
Illumina sequencing
Quantified post-PCR libraries were pooled and loaded onto an Illumina NextSeq 500/550 v2.5 flow cell for sequencing. Libraries were loaded following Illumina’s recommended loading concentrations.
Data analysis
Raw sequencing reads were downsampled to equal numbers across all samples using seqtk (v1.3-r106) and processed to remove adapter sequences and low-quality bases using trimmomatic (v0.36). Quality processed reads were aligned to the UCSC hg19 reference genome with bowtie2 (v2.3.4.1) with default parameters. Resulting SAM files were coordinate sorted using Picard SortSam (v2.18.3) and converted to BAM files with samtools view (v1.8). Duplicate reads were identified and marked from sorted BAM files with picard MarkDuplicates (v2.18.3) and used as input to collect alignment, insert size, GC bias, and various WGS metrics with Picard AlignmentSummaryMetrics (v2.18.3), Picard CollectInsertSizeMetrics (v2.18.3), Picard CollectGcBiasMetrics (v2.18.3), and Picard CollectWgsMetrics (v2.18.3), respectively. Variants were called using VarDict with a minimum of 30X coverage and a minimum of 0.5% allele frequency.
