Fig. 3: Classification of copy number alteration (CNA)-quiet and CNA-other oral cavity SCC (OCSCC) using a multicenter cohort.

a Flow chart of CNA-quiet classification. First, multiplex ligation-dependent probe amplification (MLPA) was performed on the entire cohort of 900 OCSCC as pre-screening to assign CNA-other OCSCC. CNA-quiet candidates were subjected to low coverage whole genome sequencing (lcWGS) to generate copy number (CN) plots as well as to calculate the fraction genome altered (FGA) by dividing the length of the called segments by the sum of all segment lengths and count the called segments. OCSCC with FGA < 0.20 and < 6 called segments were classified as CNA-quiet candidates, tumors with FGA ≥ 0.20 and/or ≥ 6 called segments were classified as CNA-other, and some samples had to be excluded for quality control (QC). With the remaining CNA-quiet candidates, targeted sequencing was performed to confirm presence of tumor cells, adjust the definitive CN profile, and for QC. In addition, targeted sequencing was executed to compare the mutational profiles of CNA-quiet and CNA-other OCSCC. b, c Representative CN plots of (b) CNA-quiet and (c) CNA-other OCSCCs. Segments are shown in gray and highlighted in red when called. Tumor purity was determined by the copy number changes using ACE¹⁰ and/or on the basis of variant allele frequency of somatic variants as described in the methods. d, e Frequency of gains (in blue) and losses (in red) analyzed by lcWGS for (d) 73 CNA-quiet and (e) 364 CNA-other OCSCCs. Source data are provided with this paper.