Fig. 2: Defining a cut-off to categorize copy number alteration (CNA)-quiet tumors using the cancer genome atlas (TCGA).
From: Hallmarks of a genomically distinct subclass of head and neck cancer
a Fraction genome altered (FGA) provided by cBioPortal9 plotted against purity provided by ABSOLUTE11 for 409 human papillomavirus (HPV)-negative head and neck squamous cell carcinoma (HNSCC) samples, arrow corresponds to example in (b). b Example of false copy number alteration (CNA)-quiet tumor with a provided FGA of 0.05 but a recalculated FGA of 0.48 and clearly abundant altered segments. Segments are shown in black and highlighted in red when called. c Recalculated FGA plotted against purity with arrow assigning example from (b). d Distribution of recalculated FGA across 409 HPV-negative HNSCC samples with separation of four groups: FGA < 0.08; FGA between 0.08 and 0.19; FGA between 0.20 and 0.46; and FGA > 0.46. a, c Black line: linear regression; in gray 95% confidence interval. e Percentage of CASP8 and HRAS mutations across the four CNA groups out of the total HPV-negative HNSCC TCGA cohort. p-values were obtained using a two-sided Chi-squared test. f Distribution of recalculated FGA across 409 HPV-negative HNSCC samples with separation of two groups: FGA < 0.20 and ≥ 0.20. g Called segments (y-axis) and recalculated FGA (x-axis). Tumors with FGA < 0.20 and < 20 called segments were classified as CNA-quiet and tumors with FGA ≥ 0.20 and/or ≥ 20 called segments were classified as CNA-other. h The FGA was recalculated, with correcting for stromal contamination, using R packages ACE (Poell et al.10.) + ABSOLUTE (Carter et al.11, Supplementary Fig. 2). In total, 70 tumors were assigned as CNA-quiet and 339 as CNA-other. i Relative frequency of anatomical site in CNA-quiet and -other groups. p-values were obtained using a Chi-squared test. Source data are provided with this paper.
