Mountain and Reliability plot

Fold the 5S rRNA sequence and visualize the structure. (The 5S.seq is shipped with the tutorial)

  $ RNAfold -p < 5S.seq
  $ mountain.pl 5S_dp.ps | xmgrace -pipe
  $ relplot.pl 5S_ss.ps 5S_dp.ps > 5S_ss.ps

A mountain plot is especially useful for long sequences where conventional structure drawings become terribly cluttered. It is a xy-diagram plotting the number of base pairs enclosing a sequence position versus the position. The Perl script mountain.pl transforms a dot plot into the mountain plot coordinates which can be visualized with any xy-plotting program, e.g. xmgrace.

The resulting plot shows three curves, two mountain plots derived from the MFE structure (red) and the pairing probabilities (black) and a positional entropy curve (green). Well-defined regions are identified by low entropy. By superimposing several mountain plots structures can easily be compared.

The pearl script relplot.pl adds reliability information to a RNA secondary structure plot in the form of color annotation. The script computes a well-definedness measure we call ``positional entropy'' ( $S(i) = -\sum p_{ij}\log(p_{ij})$ for those who want to know the details) and encodes it as color hue, ranging from red (low entropy, well-defined) via green to blue and violet (high entropy, ill-defined). In the example above two helices of the 5S RNA are well-defined (red) and indeed predicted correctly, the left arm is not quite correct and disordered.
For the figure above we had to rotate and mirror the structure plot, e.g.

  $ rotate_ss.pl -a 180 -m 5S_ss.ps > 5S_rot.ps

You can manually add annotation to structure drawings using the RNAplot program (for information see the man page). Here's a somewhat complicated example:

  $ RNAfold < 5S.seq > 5S.fold
  $ RNAplot --pre "76 107 82 102 GREEN BFmark 44 49 0.8 0.8 0.8 Fomark \
    1 15 8 RED omark 80 cmark 80 -0.23 -1.2 (pos80) Label 90 95 BLUE Fomark" < 5S.fold
  $ gv 5S_ss.ps

RNAplot is a very usefull tool to color plots. The -pre tag adds PostScript code to color distinct regions of your molecule. There are some predefined statements with different options for annotations listed below:

i cmark	draws circle around base i
i j c gmark	draw basepair i,j with c counter examples in grey
i j lw rgb omark	stroke segment i...j with linewidth lw and color (rgb)
i j rgb Fomark	fill segment i...j with color (rgb)
i j k l rgb BFmark	fill block between pairs i,j and k,l with color (rgb)
i dx dy (text) Label	adds a textlabel with an offset dx and dy relative to base i

Predefined color options are BLACK, RED, GREEN, BLUE, WHITE but you can also replace the value to some standard RGB code (e.g. 0 5 8 for lightblue).

To see what exactly the alternative structures of our sequence are, we need to predict suboptimal structures.