Wayne K. Dawson
WARNING: These settings should not be changed unless there is a
very good reason to do so. The default value appears to yield
reasonable results for most of the RNA currently tested. If you know
that your RNA has long purine segments (such as the hammer head
ribozyme), then you may have good reason to extend the length of the
free strand search distance. If you have some structure with a lot of
cug repeats, perhaps you will improve your predictions by increasing
the number of contiguous stems.
Number of contiguous stems (unit: number)
A contiguous stem is a set of stems that are linked together by short internal loops or short bulges. Such internal loops usually consist of 1x1, 1x2, 2x1, or 2x2 lengths and normally, the distance between two stems should be less than the persistence length. However, because the Turner energy rules artificially eliminate many types of non-WC pairs, many forms of favorable stacking that probably should be found are not. These structure may actually be essentially "stems", but because of the artificiality of such "rules", they are called loops. This approach is somewhat of a work around on this.
In addition, even when perhaps we can actually call some unpaired bases "loops", there is still the fact that when they are in close proximity, the independence of the different stems is reduced, and such structures should not be treated as strongly independent in the calculation.
This parameter decides the number of contiguous stems to permit and is
intended to limit the amount of recursive backtracking in the
Average free strand length (unit: nt)
This parameter expresses the maximum length of the free strands that is searched as though the it were part of a contiguous stem. This parameter helps set limits on how far one can extend between one stem and the next before the stems are considered independent. By chosing a large value, you ensure that a large distance can separate the stems before they become independent. However, this can only be meaningful if you know that you have Mg localizing in the purine rich regions of the structure or you have a fair number of contiguous GA pairs (or some other viable purine combination).
If you think Mg is binding in these regions, you should also invoke the Mg binding option because the combination of Mg, and purine interactions appears to be a factor in stablizing such free strands. There are a few examples of free strands that are longer than 8 nt on a side, and contain generous amounts of purines along the sequence (such as the hairpin ribozyme), however, such very long regions are rare in most RNA. Moreover, chosing a large value will slow down your calculation significantly because many more recursive paths must be searched. Since most of them will not yield much of anything different, you should not use a large value unless you know you should.