next up previous   Next: Acknowledgements Up: 6. Notes Previous: 6.1.3 Other errors

6.2 Forcing base pairs

The use of bonus energies to force base pair formation creates several problems. The bonus energy value of -500 (-50 kcal/mol) will not create a problem if only a few base pairs are being forced. Note that each closed excision is equivalent to forcing a single base pair formation (the ends of the excised fragment). However, if one wishes to fold the entire 16S rRNA from E. coli, for example, forcing all the phylogenetically determined base pairs to form, then there will be a problem. The folding program uses 16 bit integers, and the maximum absolute energy that can be stored is roughly 3,200 kcal. This will be exceeded if hundreds of base pairs are forced with a bonus energy of -50 kcal/mol per base pair. In such a case, the solution is to set the bonus energy (4.1) to -2 or -3 kcal/mol; enough so that all the desired base pairs form.

In the energy dot plot, only points corresponding to forced bases or base pairs will appear as long as the energy increment is no larger than the bonus energy. As soon as the bonus energy threshold is passed, the optimal folding, and then the suboptimal foldings will be revealed.

In N best mode, special case must be taken in choosing the ``percentage for sort'' parameter (3.5) when base pairs are forced. This is because the internal energies are distorted by the addition of bonus energies. Suppose that p is the desired ``percentage for sort'' value in a situation where the bonus energy is -b, the real optimal energy is -E and m>0 base pairs have been forced. Then the ``percentage for sort'' number p′ that must be used in this case is given by

p′ = pE + 100b

E + (m+1)b

Michael Zuker
Thu Nov 2 14:28:14 CST 1995