In silico analysis of local RNA secondary structure in influenza virus A, B and C finds evidence of widespread ordered stability but little evidence of significant covariation

Figure 1. z-score violin plots of each influenza virus, with positive (blue) and negative (red) strands for comparison. While the range of z-scores observed is broad, there is a visual trend toward the negative (structured) across all clades, with IAV segments 7 and 8 being noticeably lower than other segments. Data from ScanFold, 120 nt window, 1 nt step, 100 randomizations, 37 ˚C. Image adapted from BoxPlotR51.

Figure 2. (Left) ScanFold data for IAV 7 ( +), with motif 7_1 and 7_4 designated by a blue box. M1 and M2 versions of this transcript are illustrated for reference. The base pair track (bp track) shows arcs correlating to base pairings, where blue arcs have a < −2 score, green arcs have a < −1 z-score, yellow arcs have a < 0 z-score, and gray arcs have a z-score > 0. (Right) Motif 7_1 correlates with the published multibranch loop23, aligning with our reference sequence after global refold (gray boxes). Motif 7_4 has an extended hairpin that occludes any formation of the 5′ pairing or pseudoknot structure around the 3′ splice site (739–740, purple triangles), but a global refold reveals the published 3′ pairing25. Two sequence variations from our reference sequence are annotated at 136 and 655. ScanFold z-scores are overlaid in each nucleotide circle, with blue designating < −2 z-score. Structural images were adapted from VARNA, and the genome illustrations were adapted from NCBI.

Figure 3. ScanFold analysis for motifs (top to bottom panels, respectively) IAV ( +) 4_1, IAV (−) 6_3, IBV (−) 1_3, and IBV (−) 4_2 (locations designated with blue boxes). The base pair track (bp track) shows arcs correlating to base pairings, where blue arcs have a < −2 score, green arcs have a < −1 z-score, yellow arcs have a < 0 z-score, and gray arcs have a z-score > 0. The R-Scape calculations showed observed base pair covariance (highlighted in green), but the number of observe covarying pairs fell below the expected value (given the sequence alignment). ScanFold per nt zavg are overlaid in each nucleotide circle, with blue designating < −2 zavg. Structural images were adapted from VARNA, and the genome illustrations were adapted from NCBI.

Figure 4. ScanFold analysis for motifs ICV ( +) 6_2 and 6_4 (blue boxes, upper panel) and 5_7 (lower panel). The base pair track (bp track) shows arcs correlating to base pairings, where blue arcs have a < −2 score, green arcs have a < −1 z-score, yellow arcs have a < 0 z-score, and gray arcs have a z-score > 0. The R-Scape results for ICV ( +) 6_2 had 4 bp observed to covary (highlighted in green) with 0.0 ± 0.1 bp expected, while 6_4 had 1 observed bp with 0.0 ± 0.0 bp expected. ICV ( +) 5_7 showed a single observed bp when 0.0 ± 0.2 were expected. ScanFold per nt zavg are overlaid in each nucleotide circle, with blue designating < −2 zavg. Structural images were adapted from VARNA, and the genome illustrations were adapted from NCBI.

Figure 5. A receiver operating characteristic (ROC) analysis comparing in silico ScanFold −1 z-score predicted structures to DMS-MaPseq data for IAV (H1N1) segments 1–8. The true positive rate (TPR) is shown on the y-axis and the false positive rate (FPR) on the x-axis. Each segment is shown with a unique color and data point marker: segment 1, light blue and a diamond; segment 2, orange and an asterisk; segment 3, grey and an x; segment 4, yellow and a square; segment 5, dark blue and a circle; segment 6, green and a triangle; segment 7, purple and a cross; segment 8, maroon and a dash. The associated area under the curve (AUC) is shown below each segment in the legend.