Fig. 2

An illustration for explaining how to create a binary vector from an example RNA sequence. A moving window with 2d+1 cells (d=1 in this example) is shown in the ‘Moving window’ row. The range that the window can move along one anti-diagonal is shown by shaded cells in dot plots, and a black cell in the window corresponds to p _ij used in the main text. Note that an asterisk in the dot plot indicates a high probability of forming a base pair at the corresponding position, and the other cells without asterisk show zero probability. The window starts moving at the upper left corner on the diagonal and moves to the anti-diagonal way, which is indicated by the arrow shown in the ‘Moving window’ row. When all the cells of the window reaches the boundary of the dot plot, the window is moved back to the start position of this anti-diagonal scan and slidden by one lower right diagonal step. Note that at this moment the shape of the window is changed, e.g. the inverse ‘L’ shape into the ‘L’ shape. Repeat this scan until the window crosses over the lower right corner of the dot plot. Bold digits in the ‘Binary vector’ row are also embedded in the ‘Final binary vector,’ which are shown by bold digits in the calculation order