Adapting clinical chemistry plasma as a source for liquid biopsies
- Department of Pathology, School of Medicine, Stanford University, United States
- Department of Molecular and Medical Genetics, Oregon Health & Science University, United States
- University of British Columbia, Canada
- Clinical Laboratories, Stanford Health Care, United States
- Department of Laboratory Medicine, School of Medicine, University of California, San Francisco, United States
- Department of Medicine, Division of Infectious Diseases and Geographic Medicine, School of Medicine, Stanford University, United States
Figures
Figure 1 with 2 supplements
Overall schematic and Healthy Cohort’s comparative performance.
(a) Schematic illustration of the experimental design across two cohorts. The Healthy Cohort was a controlled experiment where plasma samples from healthy volunteer donors were collected freshly in EDTA, Streck, and heparin separators. The Hospital Cohort plasma samples were retrospectively collected in EDTA and heparin separators as part of patient care and leftover from routine clinical testing. The Hospital Cohort samples underwent an initial soft spin during the routine clinical workflow and then a hard spin after refrigerated storage. Processed samples were analyzed using whole-genome sequencing (WGS) and/or FLEXseq methylation testing. (b) Healthy Cohort fragment size distribution of cell-free DNA (cfDNA) collected across the three tubes. (c) Healthy Cohort comparison of end motif rankings in heparin separator samples versus EDTA or Streck tubes. (d) Healthy Cohort heatmap and correlation of methylation beta values in paired samples from a healthy donor (P140). Heparin separators (y-axis) are compared to EDTA and Streck tubes (x-axis). (e) Healthy Cohort estimation of cell-type proportions using methylation deconvolution for each collection tube.
Figure 1—figure supplement 1
Genome-wide coverage comparison between tube types.
(a) Genome-wide mean sequencing depth across autosomes (Chr1–22) for pooled EDTA plasma samples (n=5). Mean depth was calculated in 1 Mb bins and normalized to z-scores. (b) Genome-wide mean sequencing depth across autosomes (Chr1–22) for pooled heparin separator plasma samples (n=5). (c) Bin-by-bin comparison of z-scored mean depth between pooled EDTA and pooled heparin separator plasma across all 1 Mb bins on Chr1–22. Each point represents one genomic bin. The dashed line indicates the identity line (y=x).
Figure 1—figure supplement 2
Healthy Cohort methylation correlation and cell-type deconvolution.
Heatmap and correlation of methylation beta values with cell-type deconvolution results in paired samples from five healthy donors.
Figure 2
Degradation evaluation in the Clinical-Handling Simulation Study.
(a) Schematic of the Clinical-Handling Simulation Study. Whole blood from a healthy donor was collected in EDTA tubes and heparin separator tubes and processed either immediately (ice, 0 min; double-centrifuged plasma) or after defined delays under different temperature conditions. Plasma was generated by a soft spin followed by a hard spin, including short delays at 4°C or room temperature, extended storage at 4°C for 7 days, and a stress condition at 37°C for 24 hr. (b) Cell-free DNA (cfDNA) fragment size distributions across all conditions, shown as fraction of fragments versus fragment size. Condition codes denote tube type and handling condition, where E indicates EDTA, H indicates heparin separator, 4 indicates 4°C, RT indicates room temperature, 37 indicates 37°C, and T indicates the incubation time in hours, with 0 indicating immediate processing.
Figure 3
Viral load correlation of the Hospital Cohort.
Correlation of plasma viral DNA reads detected by whole-genome sequencing in paired cases (n=38) collected in EDTA and heparin separators. Samples were collected at two institutions (Stanford – blue, UCSF – red, EBV – circle, HHV6 – triangle, other virus – square). EBV: Epstein-Barr virus; HHV6: human herpesvirus 6; RPM: reads per million human-aligned reads.
Figure 4 with 1 supplement
Copy number analysis of the Hospital Cohort.
(a) An example of genome-wide copy number profiles from Patient 34. Case P171 was collected in a heparin separator; Case P187 was collected in an EDTA tube. (b) Boxplot of correlation of log2 copy ratio between tube types in all cases. (c) Comparison of tumor fraction in paired cases (n=32 pairs) inferred from copy number analysis.
Figure 4—figure supplement 1
CNVkit heatmap across tube types.
Genome-wide copy number profiles reanalyzed using CNVkit as a secondary CNV caller. Heatmaps show log2 copy ratio across autosomes for matched EDTA and heparin separator plasma samples. Red boxes highlight CNV-positive cases.
Figure 5 with 1 supplement
Methylation correlation and cell-type deconvolution analysis in the Hospital Cohort.
(a) Heatmap showing methylation beta values in paired samples (Patient 31, P185 versus P168). (b). Estimated cell-type proportions in the paired samples (P185 versus P168).
Figure 5—figure supplement 1
Methylation correlation across Hospital Cohort cases.
Heatmaps of methylation beta values in paired samples collected from 11 virus-positive patients in the Hospital Cohort. One additional case is shown in Figure 5.
Figure 6 with 1 supplement
Correlation of DNA quantity.
(a) Scatter plot of the normalized extracted DNA concentration based on Qubit reading. Data is from 20 individuals of the Hospital Cohort and 5 individuals from the Healthy Cohort using heparin separators and EDTA tubes. Normalized DNA concentration (ng/mL plasma) in heparin separator samples (y-axis) versus EDTA tube samples (x-axis), quantified by dsDNA Qubit and normalized to the input volume of plasma. (b) Scatter plot of the log-scale normalized concentration (Qubit readout) in the heparin separators (y-axis) versus the EDTA tube samples (x-axis). (c) Scatter plot of the ratio of human reads to a constant lambda phage DNA spike-in detected in the heparin separator sample (y-axis) versus the EDTA tube samples (x-axis). This spike-in measurement is proportional to the addressable DNA quantity of the sample.
Figure 6—figure supplement 1
Fragmentomic patterns of the Hospital Cohort.
(a) Average cell-free DNA (cfDNA) fragment size distribution collected in EDTA tubes and heparin separators from Stanford cases (n=32). (b) Average end motif frequency comparison between EDTA tubes and heparin separators, highlighting the top six motifs: CCCA, CCAG, CCTG, CCCT, CCAT, and CAAA.
Additional files
-
Supplementary file 1
Sample list of all cases included in this study.
- https://cdn.elifesciences.org/articles/108708/elife-108708-supp1-v1.xlsx
-
Supplementary file 2
Delayed-processing conditions in the clinical-handling simulation.
- https://cdn.elifesciences.org/articles/108708/elife-108708-supp2-v1.xlsx
-
Supplementary file 3
Copy number analysis across Hospital Cohort cases.
- https://cdn.elifesciences.org/articles/108708/elife-108708-supp3-v1.pdf
-
MDAR checklist
- https://cdn.elifesciences.org/articles/108708/elife-108708-mdarchecklist1-v1.pdf
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Adapting clinical chemistry plasma as a source for liquid biopsies
eLife 14:RP108708.
https://doi.org/10.7554/eLife.108708.4
