Fig. 1.

CMs and the TSM process. (A) Inference of ancestral sequences (dashed lines) for the internal nodes of a phylogenetic tree allows for locating mutations (stars) in the multiple sequence alignment (MSA) to specific tree branches. Initial and restoring mutations happening in the same branch (red and blue) appear as instantaneous CMs, and the perfect stem base pairing is retained. If the initial and restoring mutations appear in different branches (yellow and green), the mechanism triggering the CM can be studied. (B) A hypothetical sequence forming an RNA stem structure with a noncanonical G–U pair. (C–F) The Watson–Crick pairing can be corrected by DNA replication (solid arrows; first red) briefly switching to another template (blue arrows) and then returning to the original strand (green arrows). The outcome of the process (Right) depends on the relative positions of the four switch points, points 1 to 4. (C) In intrastrand switch, the newly synthesized sequence (in red) forms a hairpin and templates the replication. (D) In interstrand switch, the complementary strand templates the replication. A backward jump (point 2 < point 1) corrects the mismatch with a G–C pair (yellow circle) and cannot cause associated mutations within the loop. (E) Interstrand switch forward (point 1 < point 2) corrects the mismatch with an A–U pair and depending on positions of the switch points, may cause associated mutations. When the source sequence (in blue) is within the stem region, only the mismatch is fixed. (F) Source sequence covering the loop region inverts the loop sequence, causing multiple parallel changes (yellow asterisks). Dots and parentheses notation shows predicted secondary structures. Opening and closing parentheses indicate pairing bases in the stem; unpaired bases are marked with dots.