Table 1.
Patient and sample information.
Fig 1.
cfDNA fragment size distribution and coverage uniformity.
(A) Fragment size distribution of cfDNA extracted from plasma or the liquid portion of body fluid effusion samples. Vertical dashed lines mark local maxima. (B) Cumulative plot of percentage of nucleotides covered by a specific depth. (C) Cumulative plot of percentage of nucleotides covered by a specific depth after normalization to global mean depth of coverage of that sample. (D) Percentage of nucleotides showing no coverage genome wide and in promoter regions, 5’ UTR, exons, introns, and 3’ UTR. Annotation of the hg19 reference genome was obtained from the UCSC table browser. Genomic Trim represented hg19 reference genome after all “N” nucleotides was removed. Each bar represents the mean of 5 biological replicates for blood and tumor samples and 6 biological replicates for cfDNA samples. Error bars represents standard error. No changes between groups were statistically significant as tested by one-way ANOVA.
Fig 2.
Hierarchical clustering and PCA of the fraction of nucleotides covered in each 10kbp window.
(A) Unsupervised hierarchical clustering using the fraction of nucleotides covered in each consecutive, non-overlapping 10kbp window. Coloration of each leaf was based on sample types: red for blood, blue for tumor, orange for plasma cfDNA, and black for body fluid effusion cfDNA samples. Number in brackets in each leaf indicated sample mean coverage depth. (B) Unsupervised hierarchical clustering using the fraction of nucleotides covered in each consecutive, non-overlapping 10kbp window, with data from chromosome X and Y removed. Coloration of each leaf is based on sample types: red for blood, blue for tumor, orange for plasma cfDNA, and black for body fluid effusion cfDNA samples. Numbers in brackets in each leaf indicate sample mean coverage depth. (C) PCA on the fraction of nucleotides covered in each consecutive, non-overlapping 10kbp window with chromosome X and Y data removed. Features were not scaled to equal variance.
Fig 3.
Construction of the CNV profile of each patient.
(A) Circos plots showing the CNV profile of each patient. The outermost ring plots the ideogram of autosome 1–22 and chromosome X and Y of the hg19 reference genome. Blue segments on the ideograms correspond to unmappable regions of hg19. Each ring plots the log2 ratio in the consecutive, non-overlapping 10kbp window of a sample. Log2 ratios for blood samples were calculated against publicly available NA18535 WGS results. Log2 ratios for cfDNA, tumor, and body fluid samples were calculated against their respective blood sample. (B) Segmented CNV profiles of selected chromosomes. The x-axis represented the 10kbp window index. The y-axis represented log2 ratio. The grey horizontal line marks the theoretical 2 copy state. The red line represented the calculated copy number status of that segment.
Fig 4.
Fragment boundaries inherit nucleosome footprints.
(A-C) Plot of 5’ and 3’ fragment boundary frequencies in consecutive non-overlapping 5bp windows within ±1000bp of the TSS (A) and ±500bp of the ESS (B) and the EES (C) in blood samples. (D-F) Plot of 5’ and 3’ fragment boundary frequencies in consecutive non-overlapping 5bp windows within ±1000bp of the TSS (D) and ±500bp of the ESS (E) and the EES (F) in tumor samples.(G-I) Plot of 5’ and 3’ fragment boundary frequencies in consecutive non-overlapping 5bp windows within ±1000bp of the TSS (G) and ±500bp of the ESS (H) and the EES (I) in cfDNA samples.
Fig 5.
Nucleosome footprints interfered with cfDNA WGS sequencing uniformity.
(A-C) Coverage depth at each nucleotide normalized to regional mean depth within ±1000bp of the TSS (A) and ±500bp of the ESS (B) and EES (C) in blood samples. Each grey line represents a sample. Red line represented the mean value of all samples. (D-F) Coverage depth at each nucleotide normalized to regional mean depth within ±1000bp of the TSS (D) and ±500bp of the ESS (E) and EES (F) in cfDNA samples. Each grey line represents a plasma cfDNA sample. Black lines represents the body fluid cfDNA sample. Orange line represents the mean value of all plasma cfDNA samples. Vertical dashed lines mark local coverage depth maxima. (G-I) Coverage depth at each nucleotide normalized to regional mean depth within ±1000bp of the TSS (G) and ±500bp of the ESS (H) and EES (I) in tumor samples. Each grey line represents a sample. Blue lines represent the mean value of all samples. (J-L) Per base GC content within ±1000bp of the TSS (J) and ±500bp of the ESS (K) and the EES (L).