Measurement Biases Distort Cell-Free DNA Fragmentation Profiles and Define the Sensitivity of Metagenomic Cell-Free DNA Sequencing Assays.
Author(s): Chang A, Mzava O, Lenz JS, Cheng AP, Burnham P, Motley ST, Bennett C, Connelly JT, Dadhania DM, Suthanthiran M, Lee JR, Steadman A, De Vlaminck I
Publication: Clin Chem, 2021, Vol. , Page
PubMed ID: 34718476 PubMed Review Paper? No
Purpose of Paper
The purpose of this paper was to investigate potential effects of cell-free DNA (cfDNA) isolation and library preparation methods on the observed size profile and microbial fraction of cfDNA from urine.
Conclusion of Paper
Chromosomal and mitochondrial cfDNA fragment length distributions were dependent on cfDNA extraction method, but microbial cfDNA length was comparable among the methods examined. While the measured plasma cfDNA length was similar to that computationally expected, for urinary cfDNA there was a large discrepancy. Importantly, while 43.4% of chromosomal and 34.0% of mitochondrial cfDNA was >100 bp when isolated using the MagMax isolation kit, 11.7% of chromosomal and 8.7% of mitochondrial cfDNA fragments are expected to be >100 bp. Further, the 167 bp nucleosome peak was not observed when values were not corrected for isolation bias. The authors state that applying the bias correction to the plasma cfDNA profile allowed for detection of the 308 bp dinucleosome peak. Because of this bias, the fraction of microbial DNA was up to 6-fold higher when cfDNA was isolated using the Urine Cell-Free Circulating DNA Purification Kit, which favors short cfDNA, in comparison to the CNA Kit. Length bias in cfDNA library preparation was not predictive of the fraction of microbial DNA, but single strand-based library preparation methods outperformed double strand-based methods like the NEBNext Ultra II DNA Library Prep Kit.
Studies
-
Study Purpose
The purpose of this study was to investigate potential effects of cfDNA isolation and library preparation methods on the observed size profile and microbial fraction of cfDNA from urine. Midstream urine specimens collected from 83 kidney transplant patients were centrifuged at 3000 x g for 30 min, aliquoted and stored at -80°C until analysis. Midstream urine specimens collected from 66 patients diagnosed with tuberculosis were mixed with Streck cell-free DNA urine preservative, centrifuged at 3000 x g for 30 min, aliquoted and stored at -80°C until analysis. Midstream urine specimens were collected from 40 patients diagnosed with a urinary tract infection (processing and storage not specified). EDTA blood from 6 patients receiving double-lung transplants was centrifuged at 16,000 x g for 10 min and plasma was stored at -80°C, within 24 h of blood collection. To investigate potential effects of cfDNA isolation method, cfDNA was isolated from urine of 30 kidney transplant patients and 66 patients with tuberculosis using the QIAamp Circulating Nucleic Acid Kit, Urine Cell-Free Circulating DNA Purification Kit, QIAquick Nucleotide Removal Kit, MagMAX Cell-Free DNA Isolation Kit, and a modification of the MagMAX Cell-Free DNA Isolation Kit. Potential effects of cfDNA isolation method on the observed fraction of microbial cfDNA was investigated by extracting cfDNA using each of the aforementioned methods from the urine of 20 patients diagnosed with a UTI. Potential effects associated with library sequencing method on the observed fraction of microbial cfDNA was investigated using urine from 10 patients with a UTI by constructing matched libraries using the SRSLY NGS Library Prep Kit, the NEBNext Ultra II DNA Library Prep Kit for Illumina, and the Meyer method. Unless otherwise specified, cfDNA was extracted with the QIAamp Circulating Nucleic Acid Kit and sequencing libraries were generated using the Meyer method. Libraries were pair-end sequenced on a NextSeq 500 platform. The fragment length density profile was obtained using the hist function from the R graphics package.
Summary of Findings:
Chromosomal and mitochondrial cfDNA fragment length distributions were dependent on cfDNA extraction method, but microbial cfDNA length was comparable among the methods examined. While the measured plasma cfDNA length was similar to that computationally expected, for urinary cfDNA there was a large discrepancy. Importantly, while 43.4% of chromosomal and 34.0% of mitochondrial cfDNA was >100 bp when isolated using the MagMax isolation kit, 11.7% of chromosomal and 8.7% of mitochondrial cfDNA fragments are expected to be >100 bp. Further, the 167 bp nucleosome peak was not observed when values were not corrected for isolation bias. The authors state that applying the bias correction to the plasma cfDNA profile allowed for detection of the 308 bp dinucleosome peak. Because of this bias, the fraction of microbial DNA was up to 6-fold higher when cfDNA was isolated using the Urine Cell-Free Circulating DNA Purification Kit, which favors short cfDNA, in comparison to the CNA Kit. Length bias in cfDNA library preparation was not predictive of the fraction of microbial DNA, but single strand-based library preparation methods outperformed double strand-based methods like the NEBNext Ultra II DNA Library Prep Kit.
Biospecimens
Preservative Types
- Frozen
Diagnoses:
- Other diagnoses
Platform:
Analyte Technology Platform DNA Next generation sequencing Pre-analytical Factors:
Classification Pre-analytical Factor Value(s) Next generation sequencing Specific Technology platform SRSLY NGS Library Prep Kit
NEBNext Ultra II DNA Library Prep Kit for Illumina
Meyer method
Analyte Extraction and Purification Analyte isolation method QIAamp Circulating Nucleic Acid Kit
Urine Cell-Free Circulating DNA Purification Kit
QIAquick Nucleotide Removal Kit
MagMAX Cell-Free DNA Isolation Kit
Modified MagMAX Cell-Free DNA Isolation Kit