Circulating cell-free DNA (cfDNA) serves as a non-invasive biomarker within plasma, providing insights into the genetic characteristics of various diseases. In this study we have evaluated the performance of some commonly used blood collection tubes and demonstrate how the choice of collection tube and sample storage time can impact sequencing performance.

Methods and Results

Donor blood was collected, aliquoted in volumes of 10 ml into five types of collection tubes, and shipped overnight at 4°C to our facility. To simulate a range of real-world processing delays, the samples were stored at 4°C and extracted at day 0 (day of receipt), day 3 and day 6 using K2 EDTA, K3 EDTA, Sodim Citrate and cfDNA Streck tubes. Samples collected in heparin-based storage tubes were processed at day 3 only.

cfDNA was extracted and prepared for sequencing following the sequencing of human cfDNA with the Ligation Sequencing Kit V14 (SQK-LSK114) protocol.

Briefly, blood was centrifuged, allowing for the aspiration of approximately 4.5 ml of plasma per tube. After a second centrifugation to pellet residual cells, 4 ml of plasma was extracted per sample. cfDNA was isolated using the QIAamp MinElute ccfDNA Midi Kit (QIAGEN), following the manufacturer's instructions. The extracted cfDNA was quantified using a Qubit dsDNA HS Kit (Thermo Fisher) and fragment length assessed using the Femto Pulse (Agilent) (Table 1 and Figure 1). Libraries were prepared with the Ligation Sequencing Kit V14 (SQK-LSK114) and sequenced on PromethION R10.4.1 Flow Cells for 72 hours, setting the minimum read length to 20 bp.

Read lengths and flow cell outputs were recorded and documented (Table 1 and Figure 2). Assessment of read lengths was performed by categorising reads on the basis of length. Each size bin corresponds to a specific nucleosomal peak, up to 1100 bases, with read length > 1100 bases categorised in a separate bin (as illustrated in Figure 3A).

Table 1. Library preparation and sequencing statistics for cfDNA libraries prepared in this study. The volume of plasma isolated from each collection tube consistently measured ~ 4.5 ml, allowing 4 ml of spun down plasma to be used as input into extraction. At Day 0, DNA yields and read N50 vales were comparable between tubes. However, after a period of storage, DNA yields and read N50 values were notably higher from K2/K3 EDTA and heparin collection tubes compared to those from Sodium Citrate or Streck tubes: it is postulated that this is likely a consequence of gDNA contamination occurring when using these storage tubes.

Figure 1. Fragment length profiles measured using the Femto Pulse system (Agilent). Fragment length profiles look equivalent at Day 0, but longer fragments appear after a period of storage, although this appears less significant for sample collected in cfDNA Streck tubes.

Figure 2. Read length profiles measured using Oxford Nanopore Technologies sequencing. The characteristic nucleosomal peak pattern was present in all datasets, except for those from heparin storage tubes, where read lengths differed significantly from typical cfDNA samples.

Figure 3. (A) Sequence length profile of extracted cfDNA showing the distribution and binning of nucleosome peaks 1-5 & > 5. Relative (B) and total (C) amount of data generated from each distinct nucleosomal peak 1-5, or where read length is > 1100 bases (> 5) (as defined in Figure 3). Data from sequences longer than 5 nucleosomal peaks is highlighted in red, serving as an indicator of potential gDNA contamination (which is typically longer than cfDNA molecules). At Day 0, the proportion data generated from longer DNA reads is low and equivalent across the tube types. However, by day 3 & 6 of sample storage, K2/K3 EDTA tubes (and Sodium Citrate by day 6) generate a larger proportion of read lengths longer than 1100 bases compared to Streck (Panel A). Despite this however, it should be noted that the absolute amount of data coming specifically from reads lengths that correspond to nucleosomal peaks 1-5 does not appear to be significantly affected by the choice of collection tube (Panel B).

Summary

We observed that when samples are extracted on the day of receipt (Day 0), there is little difference in the yield of material or sequencing performance of samples collected in K2 EDTA, K3 EDTA, Sodium Citrate or cfDNA Streck tubes. However, when samples are stored, rather than extracted when fresh, K2 EDTA and K3 EDTA consistently produced longer DNA fragments and sequencing read lengths compared to cfDNA Streck tubes. Although these longer fragments might suggest genomic DNA (gDNA) contamination, this interpretation is not conclusively proven and despite the presence of longer fragments in K2 EDTA and K3 EDTA, the absolute output from reads corresponding to cfDNA nucleosomal peaks appears to be unaffected by tube type (Figure 3C).

We advise customers against the use of Heparin storage collection tubes, as the data suggest the formation of cfDNA polymers which is consistent with the observations reported by de Vries et al. (2019).

References

de Vries, J.C., Barendrecht, A.D., Clark, C.C. et al. Heparin Forms Polymers with Cell-free DNA Which Elongate Under Shear in Flowing Blood. Sci Rep 9, 18316 (2019). https://doi.org/10.1038/s41598-019-54818-3

Change log

Version	Change
v1, August 2024	Initial publication