Skip to main content

Evaluation of variability in cell-free DNA extraction efficiency from plasma and urine and spike-in normalization.

Sandberg F, Kueng N, Largiadèr CR, Amstutz U

Publication: Sci Rep (2025) • Vol 15 • Page 22999PMID: 40596083View on PubMed

Purpose

This paper compared different methods to extract cell-free DNA (cfDNA) from urine and investigated sources of variability during cfDNA extraction from the urine and plasma of healthy volunteers. 

Conclusion

Following extraction using Q-sepharose, the cfDNA yield and number of AGO1 copies were very strongly correlated with the urine input volume. The size distribution pattern of the urinary cfDNA was not correlated with the urine input volume, as it differed between Q-sepharose and the Zymo Quick-DNA Urine Kit (Zymo) extraction methods. Following extraction with Q-sepharose, there were more fragments < 90 bp, less high molecular weight DNA (> 1,000 bp), and more fragments that were 89 bp in length relative to those that were 180 bp than when extraction was with the Zymo Kit. However, the mean cfDNA concentration was higher when extraction was with the Zymo Kit relative to the Q-sepharose kit (3.72 versus 2.85 ng/mL urine). When potential sources of variability were investigated using a spiked-in 180 bp fragment of CEREBIS that was extracted and assayed in triplicate from the urine of three men and three women, the largest variability occurred between specimens (96.17% with Q-sepharose extraction and 89.03% with Zymo extraction), followed by between males and females (3.59% for Q-sepharose extractions and 10.92% for Zymo extractions), with only negligible amounts of variability found inter- or intra- extraction (<0.5% all). Similarly, when the urine was spiked with both 89 and 180 bp fragments of CEREBIS, the majority of the variability occurred between individuals (95.51% with Q-sepharose extraction and 86.23% with Zymo extraction), but while variability between males and females continued to have a larger effect than intra-extraction variability when extraction was with Zymo (14.49% versus 0.27%), inter-extraction variability (not assessed with Zymo) was higher than variability between specimens from men and women when extraction was with Q-sepharose (5.25% versus 0.00%). Importantly, mean extraction efficiencies for isolation of the 180 bp and 89 bp spiked-in DNA were higher using the Zymo Kit (61.12% and 64.53%, respectively) than Q-Sepharose Kit (28.11 and 38.69%, respectively). Correction of the cfDNA yield with either the extraction efficiency of the 89 or 180 bp fragment of CEREBIS decreased the variance when extraction was with Q-sepharose but not the Zymo Kit. However, correction using creatinine provided the greatest reduction in variability regardless of the extraction method.

Notably, the cfDNA pattern in plasma was distinct from that in urine. When cfDNA was extracted from triplicate aliquots of pooled plasma (from three men) by three different researchers on three different days and quantified by ddPCR in triplicate, the largest variance in all three assays was within a single extraction (assay variability, 46.09-66.86%), followed by between extractions by the same researcher on the same day (23.45- 28.48%) and extractions by the same researcher on different days (9.70-25.63%) with the least variability attributed to the researcher performing the extraction (0-4.53%). When urine from 3 men was spiked with a 180 bp fragment of CEREBIS and then extracted in triplicate, 84.55% of the variability was found to be inter-individual, with 6.76% and 8.68% of the variability attributed to intra- (assay) and inter- extraction, respectively. When the cfDNA copies were corrected for extraction efficiency using the spike-in control (180 bp CEREBIS), the variability in the assay was reduced, and the inter-extraction variability in SRY and RPP30, but not AGO1, was reduced.

Studies (2)

Purpose

This study compared two different methods to extract cell-free DNA from urine and investigated the variability associated with the individual, sex, the researcher performing the extraction, the extraction itself and the assay. Additionally, the effect of urine volume on cfDNA yield and the number of copies of AGO1 were investigated. Urine was collected from three healthy men and three healthy women into LoBind tubes containing Streck Urine Preserve on two different days. Urine was centrifuged at 2000 g for 15 min and the supernatant was stored at -80°C. Creatinine was quantified in each urine specimen. Some urine from the three men and three women were pooled (a pool of urine from males and a separate pool of urine from females), aliquoted, spiked with 85,000 copies of a 180 bp fragment of CERBIS, and cfDNA extraction was performed in triplicate on three separate days by three different researchers and then assayed in triplicate. To compare intra- versus inter-individual variability, the remaining individual specimens were spiked with 45,000 copies each of a 180 bp and 89 bp fragment of CERBIS, and the cfDNA was extracted in triplicate using both kits and then assayed in triplicate. cfDNA was extracted from urine using the Zymo Quick-DNA Urine Kit or an in-house method using Q Sepharose Fast Flow and Micro Bio-Spin Chromatography Columns, and clean-up was with QIAquick Spin Columns. cfDNA was quantified using the Qubit 1X dsDNA HS Assay Kit and by droplet digital PCR (ddPCR) amplification of AGO1. The size distribution of the isolated cfDNA was analyzed using the High Sensitivity DNA Kit with the 2100 Bioanalyzer system.

Summary of Findings

Following extraction using Q-sepharose, the cfDNA yield and number of AGO1 copies were correlated with the urine input volume (r=0.995, p<0.001 and r=0.999, p<0.001, respectively). The size distribution pattern of the urinary cfDNA was not correlated with the urine input volume as itdiffered between Q-sepharose and the Zymo Quick-DNA Urine Kit (Zymo) extraction methods. Following extraction with Q-sepharose, there were more fragments < 90 bp, less high molecular weight DNA (> 1,000 bp), and more fragments that were 89 bp in length relative to those that were 180 bp than when extraction was with the Zymo Kit. The Zymo Kit also failed to isolate the 50 bp DNA from phosphate buffered saline (PBS) spiked with a DNA ladder. However, mean cfDNA concentration was higher when extracted with the Zymo Kit relative to the Q-sepharose method (3.72 versus 2.85 ng/mL urine). When potential sources of variability were investigated using a spiked-in 180 bp fragment of CEREBIS that was extracted and assayed in triplicate from the urine of three men and three women, the largest variability was between individuals (96.17% with Q-sepharose extraction and 89.03% with Zymo extraction), followed by between males and females (3.59% for Q-sepharose extractions and 10.92% for Zymo extractions), with only negligible amounts of variability associated with inter- or intra- extraction (<0.5% all). Similarly, when the urine was spiked with both 89 and 180 bp fragments of CEREBIS, the majority of the variability occurred between individuals (95.51% with Q-sepharose extraction and 86.23% with Zymo extraction), but while variability between males and females continued to have a larger effect than intra- extraction variability when extracted with the Zymo Kit (14.49% versus 0.27%), inter-extraction variability (not assessed with Zymo Kut) was higher than the variability that occurred between specimens from men and women when extracted with Q-sepharose (5.25% versus 0.00%). Importantly, the mean extraction efficiency for isolation of the 180 bp and 89 bp spiked-in DNA were higher using the Zymo (61.12% and 64.53%, respectively) than the Q-Sepharose Kit (28.11 and 38.69%, respectively). Correction of the cfDNA yield with either the extraction efficiency of the 89 or 180 bp fragment of CEREBIS decreased the variance when extraction was with Q-sepharose, but not the Zymo, Kit. However, correction using creatinine provided the greatest reduction in variability regardless of the extraction method.

  • Biospecimens

    FluidUrine
  • Preservation Types

    Frozen
  • Diagnoses

    Normal
  • Platforms

    DNADigital PCR
    DNAFluorometry
    DNAAutomated electrophoresis/Bioanalyzer
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    PreaquisitionPatient gender
    Female
    Male
    Biospecimen Aliquots and ComponentsAliquot size/volume
    10 mL
    20 mL
    30 mL
    40 mL
    50 mL
    Analyte Extraction and PurificationAnalyte isolation method
    Isolated by three different researchers
    Q-sepharose
    Three extractions by each researcher
    Zymo Kit
    Digital PCR SpecificData handling
    Corrected for CEREBIS 180 bp
    Corrected for CEREBIS 89 bp
    Corrected for creatinine
    Digital PCR SpecificLength of gene fragment
    180 bp CEREBIS Spike-in
    89 bp CEREBIS spike-in