Figure 1.
Bioinformatics pipeline of our copy number variation analysis strategy.
Table 1.
Date production of sample using in this study.
Figure 2.
RCR distribution in correction process.
These plots represent the correction effects of the two-step correction methods in an observed sample (CS-NA18632). (a, b, c) Performance in the RCR distribution [original (a), after correction step-1 (b), after correction step-2 (c)]. (d, e, f) Performance in coefficient of variation changement [original (d), after correction step-1 (e), after correction step-2 (f)]. (g, h, i) Performance of the RCR correction effect in the regions that have the similar GC content (i.e. 38∼42%) but different genome structure (the left box is the stable regions, the right box is the regions that close to centromeres, telomeres or N regions) [original (g), after correction step-1 (h), after correction step-2 (i)].
Figure 3.
The sensitivity and specificity of PSCC in simulated CNVs detection process.
In these plots, X-axis stands for the CNV size ranged from 0.2 to 20×, and Y-axis was the percentage. The sequencing depth are color-coded and distinguished by characters. (a, b) Sensitivity of simulated CNV detection [Duplication (a), Deletion (b)]; (c, d) Specificity of simulated CNV detection [Duplication (c), Deletion (d)].
Table 2.
Statistics of CNVs detection by PSCC for 34 clinical samples.
Figure 4.
The detection results of confirmed CNVs using PSCC, SegSeq and ReadDepth.
The four plots had shown the sensitivity and specificity in three methods with a low coverage sequencing depth. (a) The sensitivity in 2× sequencing; (b) The sensitivity in 0.5× sequencing; (c) The specificity in 2× sequencing; (d) The specificity in 0.5× sequencing.