Predicting equilibrium pair probabilities

1. Run RNAfold -p -MEA to compute the partition function and pair probabilities as well as the maximum expected accurarcy.
2. Look at the generated PostScript files test_ss.ps and test_dp.ps

     $ RNAfold -p --MEA < test.seq 
     CCGCACAGCGGGCAGUGCCCA
     .........((((...)))). ( -5.70)
     ,,{{...}}||((...))}}. [ -6.25]
     ..((...))((((...)))). { -4.80 d=3.57}
     ..((...))((((...)))). { -4.80 MEA=14.90}
      frequency of mfe structure in ensemble 0.408876; ensemble diversity 4.29
   

Here the last three lines are new compared to the text output without the -p -MEA options. The partition function is a rough measure for the well-definedness of the MFE structure. The third line shows a condensed representation of the pair probabilities of each nucleotide, similar to the dot-bracket notation, followed by the ensemble free energy (-kT ln(Z)) in kcal/mol. The next two lines represent the centroid structure with its free energy, its distance to the ensemble and the MEA. The last line shows the frequency of the MFE structure in the ensemble of secondary structures and the diversity of the ensemble.

"." denotes bases that are essentially unpaired, "," weakly paired, "$\vert$"strongly paired without preference, "{},()" weakly ($>$33%) upstream (downstream) paired or strongly ($>$66%) up-/downstream paired bases, respectively.

Note that the MFE structure is adopted only with 40% probability, also the diversity is high for such a short sequence.

For rotating the secondary structure plot there is a usefull tool called rotate_ss.pl included in the ViennaRNA Package. Just read the perldoc for this tool to know how to handle the rotation and use the information to get your secondary structure in a vertical position.

  $ perldoc rotate_ss.pl