Introduction

Short Fragment Eliminator Expansion (EXP-SFE001) is used to prepare DNA samples for Oxford Nanopore Technologies library preparation and sequencing, by removal of short DNA fragments (<10 kb) from genomic DNA (gDNA) samples.

The size selection method can be found on the community by using this link: Removal of short fragments with the Short Fragment Eliminator Expansion (EXP-SFE001)

The size selection process involves mixing gDNA with an equal volume of Short Fragment Eliminator (SFE) buffer, followed by centrifugation to pellet high molecular weight (HMW) gDNA, with elution achieved by re-suspending the pellet in an aqueous solution (referred to as primary centrifugation and primary elution, respectively). The supernatant from the primary centrifugation is then re-centrifuged to elute additional gDNA (secondary centrifugation and secondary elution, respectively). DNA from both primary and secondary elutions is then combined to maximise recovery whilst ensuring effective size selection.

This document outlines the results of experiments conducted to characterise the use of the SFE buffer on gDNA extracted using various methods, sample types, and input masses.

Methods

To assess the impact of SFE buffer on a variety of DNA inputs, gDNA was extracted from two sources: HG002 cells grown in culture and human blood samples obtained from a commercial vendor. Three different extraction methods were used to provide a range of fragment length profiles: Puregene™ (QIAGEN), MagAttract™ (QIAGEN), and DNeasy™ (QIAGEN). For each extraction method, multiple input masses were tested (1, 3, 5 and 10 µg) to evaluate the effect of DNA mass on the SFE process. Recovery efficiencies were measured from the primary and secondary elutions to highlight the requirement of the secondary centrifugation/elution steps.

In addition, a separate experiment was conducted to evaluate the performance of SFE buffer at a higher buffer-to-sample ratio using gDNA extracted from saliva with the GeneFiX™ DNA Isolation Kits (Isohelix). A fixed mass of 5 µg gDNA was used in this experiment to explore whether this approach could improve recovery efficiencies.

Extracted gDNA samples were assessed for fragment size using Femtopulse™ and Genomic DNA 165 kb kit (Agilent) (Figure 1).

Figure 1. FemtoPulse analysis showing the gDNA fragment length profiles of the three extraction methods tested on gDNA extracted from cells. gDNA extracted using Puregene appeared to have the longest fragments present, followed by MagAttract. DNeasy appeared to generate the shortest fragments from the extraction.

To evaluate the recovery efficiency and fragment size distribution of SFE-treated samples, DNA concentrations were measured using a Qubit™ Fluorometer (Invitrogen). gDNA samples were sequenced on Oxford Nanopore Technologies DNA sequencing platforms to assess size selection performance.

Results

For gDNA extracted using Puregene, MagAttract and DNeasy; recovery efficiency was typically >50% for input masses of 3 µg, 5 µg, and 10 µg when combining yields from both the primary and secondary elutions. In contrast, 1 µg inputs consistently failed to achieve sufficient recovery for downstream sequencing across all sample types and gDNA extraction methods, with recovery efficiencies below 5%.

The Puregene and MagAttract extraction methods showed comparable total recovery efficiencies for both cultured cells and blood samples. The gDNA source (whether from blood or cultured cells) did not appear to significantly influence the recovery efficiency. For DNeasy, recovery efficiencies were generally slightly lower, which may be a result of the shorter fragment lengths generated using this extraction method.

The use of 3 µg or 5 µg gDNA typically resulted in low yields from the primary elution, and the majority of gDNA was recovered from the secondary elution, significantly improving overall yield and ensuring sufficient DNA recovery for sequencing. For 10 µg inputs, the secondary centrifugation step was occasionally not required (Puregene and MagAttract DNA), as recovery from the primary elution was already sufficient for sequencing (Table 1 and Figure 2).

Table 1. Performance metrics for the tested extraction methods tested with blood and cultured cell gDNA origins for various input quantities. * For 10 µg inputs, the secondary centrifugation step was omitted, meaning the total recoveries reported do not include the additional DNA that could have been obtained through this step.

Figure 2. Proportion of DNA recovered from primary and secondary centrifugation/elution for the tested extraction methods and the corresponding gDNA input origin.

gDNA extracted from saliva generally resulted in low overall recovery efficiencies, which were improved by using a higher SFE buffer-to-sample ratio (2:1 rather than 1:1) to achieve sufficient recovery for certain workflows (Table 2 and Figure 2).

Table 2. Performance metrics for the GeneFiX DNA Isolation Kit extraction method tested with saliva gDNA origins for 5 µg input and processed with 2:1 and 1:1 SFE ratios.

Size selection performance was broadly equivalent across all sample types, gDNA extraction methods, and input masses. Sequencing results showed the effective removal of short DNA fragments, and that the gDNA extraction method did not significantly influence read length distributions (Figure 3). Similarly, input mass had no significant effect on sequencing read length distributions (Figure 4). Furthermore, whether DNA was recovered from the primary or secondary elution had no significant impact on read length distributions (Figure 5). Size selection performance was found to be slightly reduced when using a 2:1 rather than a 1:1 SFE buffer-to-sample ratio (Figure 6). However, higher recoveries were achieved using the 2:1 ratio (Table 2).

Figure 3. The plot shows the effect of size selection on 10 µg of gDNA extracted from cultured cells using three gDNA extraction methods: Puregene, MagAttract, and DNeasy. The SFE buffer effectively removed short DNA fragments from all three gDNA samples, as demonstrated by the sequencing read length distributions of SFE-treated gDNA compared to the gDNA input controls.

Figure 4. This plot shows comparable effectiveness in short fragment removal for 5 µg and 10 µg DNA inputs. gDNA was extracted from cultured cells or blood using the Puregene gDNA extraction method and treated with SFE, with untreated gDNA included for comparison.

Figure 5. The plot shows equivalent SFE size selection performance by removing short fragments from gDNA extracted from cultured cells using the Puregene method, with a fixed input mass of 5 µg used. Size selection performance was consistent regardless of whether the gDNA was recovered from the primary or secondary centrifugation.

Figure 6. This plot shows a slight reduction in read length when a higher (2x) ratio of SFE buffer was used. gDNA was extracted from cultured cells or blood, and a fixed input mass of 5 µg was used in SFE size selection at a ratio of 1:1 or 2:1 SFE buffer : sample ratio.

Discussion

Our findings indicate that relying on the primary eluate is generally insufficient for obtaining adequate DNA for sequencing, except when using 10 µg of gDNA extracted with Puregene or MagAttract. To maximise yields, we incorporate the secondary centrifugation step as standard practice.

Using an input mass of at least 3 µg was critical for all gDNA types, as lower inputs (1 µg) consistently failed to produce sufficient yields for sequencing. Additionally, sequencing read length distributions were unaffected by whether DNA was recovered from the primary or secondary elutions, demonstrating that combining both is an effective approach to ensure efficient sample recovery.

We also attempted to mimic the conditions achieved during the secondary centrifugation, by doubling the initial centrifugation time to see if this could be performed instead of a separate re-centrifugation step of the supernatant. However, we found this approach did not improve recovery efficiencies (data not shown). Separating the supernatant from the DNA pellet and performing the secondary centrifugation on the supernatant was the only technique we tested that successfully led to increased recovery.

Increasing the SFE buffer-to-sample ratio can improve recovery efficiencies in certain cases. For instance, gDNA extracted from saliva typically shows low recovery from combined primary and secondary elutions when using a 1× SFE buffer ratio, and a higher buffer-to-sample ratio is necessary to achieve sufficient recovery for this specific workflow. However, this adjustment is not our default recommendation due to observed subtle detrimental effects on read length (Figure 6).

To evaluate the impact of storage duration on sequencing performance, size-selected gDNA was stored at 4 °C and -20 °C for 0 days, 14 days, and 36 days. Sequencing results indicated no significant reduction in read length, even after 62 days of storage for both 4 °C and -20 °C (data not shown).

Change log

Version	Change
v1, Jan 2025	Initial publication