Fig. 6

Junctions have the potential to generate variant open reading frames. a–c ORFs were predicted directly from transcript-derived reads for the three dRNAseq datasets. Each ORF was aligned against the protein sequences of canonical SARS-CoV-2 genes using the DIAMOND aligner. Variant ORFs were defined as ORFs that were assigned to a canonical SARS-CoV-2 protein but had an unexpected start or stop position, while perfectly aligning ORFs were considered canonical. The percentage of canonical and variant ORFs for each protein is plotted. d–l Schematics of M, N, and ORF1a are displayed with the approximate location of predicted transmembrane domains labeled in red. A histogram of the start and end sites of variant M, N, and ORF1a ORFs are displayed. Start sites of each variant ORF are on top, and end sites are on the bottom of each panel. Percentages represent the percentage of each ORF that is variant. Histogram bin sizes: M, 1; N, 10; ORF1a, 20. NTD: N-terminal domain. RBD: Receptor binding domain. CD: connector domain. NSP: nonstructural protein. m–o The identity of ORF1a-fusion partners is plotted on the Y-axis, with the count of such fusions on the X-axis. The top 10 fusion partners for each sample are represented. Color indicates if the fusion partner is on the N and C terminus of ORF1a, if the terminus is ambiguous, or if the fusion is a “self” fusion between an upstream and downstream region of ORF1a