Nucleotides on the hook

  • 1
  • Article
  • Updated 8 years ago
Sometime ago I did a post on nucleotides in the multiloop area

Back then I noticed the same tendency when adding nucleotides in the hook area, I just didn't had an oppinion of the purpose of it.

Adding nucleotides in the hook area, works much like adding nucleotides in multiloops. They can change the energy level inside multiloop or hook up and down, if wrong placed they can break up the multiloop and neck or they can do nothing.

Note: If you ad anything in the hook area, it mess up your ability to read the design in both Nupack meltplot and RNA fold. It changes the overall appearance of the design in fx. positional entropy (turns owerall red, which normally is good, but here tells you nothing) and the curve in meltplot, so you can't see if your design is good or bad.

I think nucleotide(s) should only be added in the hookarea if it changes the energy level in the hook in the prefered direction (the one that makes this specific design stronger). Which is usually higher negative numbers. I think it's the same with nucleotides in multiloop, better is often higher negative numbers.

Adding a nucleotide that does not change the energylevel inside the hook, might still get you a highscoring design. Like mine Mat with robotneck and tetraloops, 95%, which has flaws like 2 identical arms. I'm not sure if that red nucleotide in the hookarea did anything god or bad for the design. Let me hear if some of you have an oppinion about those nucleotides that does not change energylevel in hook- and multiloop area.

Ealier we were having big problems making the neck stick, which is propably why people for a period sort of didn't use nucleotides much in the hook area. Generally it have not been a good idea to do much with the hookarea. As Jeehyung says: Most people choose to use all As for multiloops and the hook area. If you ad nothing here, it can't ruin the rest of your design.

People are now getting an idea about how to add nucleotides in the hook area. I noticed something funny in Aldo's Cyborg design (Branches 100%)



This was a 100% scoring designs with opposite turning GC-pair between neck and multiloop. Here the energy is lowered in both the middle multiloop and hook area. Maybe to allow the low energy neck.

I think nucleotides in hook area (and multiloop area) are handles that can be used to control the stability of the design. They can be used to raise or lower the energy inside the hook according to what is needed for the general energy level in the design.

Uses of nucleotides in the hookarea

- To raise energy level in the hook, could ad stability to neck/hook.
- To lower energy in the hook, to allow use a low energy neck
- To center energy balance as to make the energy distribution in the design more even (just a predictive guess, as I noticed an slight skewing in energy distribution in the branches designs (see my post Even energy distribution continued)
Photo of Eli Fisker

Eli Fisker

  • 2239 Posts
  • 498 Reply Likes

Posted 8 years ago

  • 1
Photo of aldo

aldo

  • 35 Posts
  • 2 Reply Likes
People are now getting an idea about how to add nucleotides in the hook area. I noticed something funny in Aldo's Cyborg design (Branches 100%)

For the record, I just copied the neck and hook from a NUPACK design that had a successful neck. So "people" might not be the most accurate term here. :-P

Regardless, I've just resubmitted Cyborg 2 (which has the same neck and hook as Cyborg 1 and also scored 100) without the C at 4 to see what happens.
Photo of Eli Fisker

Eli Fisker

  • 2239 Posts
  • 498 Reply Likes
Good point, Aldo! People are an absolute less accurate term for robots. :) The robots have been heavy users of nucleotides in both multiloop ring and the hook area. At the beginning it seemed with bad luck, but now sometimes with luck.

I will look forward to see how this design without C af 4 does. Your experiment could answer some questions. Or start some new.
Photo of Eli Fisker

Eli Fisker

  • 2239 Posts
  • 498 Reply Likes
Now the results from Aldo's experiment got back. In his 100% design Cyborg 2, he had a green nucleotide at position 4. To test my hypotesis he removed this green nucleotide from the hook, to see if nothing would happen, like I predicted.

Results:

I was only partly right

1) Energy level did not change
2) Score result did not change

But
3) Shapedata (in the blue yellow mode) did change. That I did not forsee. This might however be the explaination why adding nucleotides in hook area, (even if not changing the energy), often makes the designs hard to read in Vienna RNA fold and Nupack.

That was originally why I disliked nucleotides in hookarea, because in the beginning of my eterna time, I was a big fan of analysis tools. :) Now I see the value of adding nucleotides in both hook area and multiloop area, mainly though if they change the energy and ad strength to the design - if needed. Adding boost points, as Iroppy would call it.

4) Which of the designs are best, now that they are not the same? Cyborg 2.1 without the green nucleotide have a more greyish shade than blue, when viewing the shapedata. Cyborg 2 have a more blue shade. Does this mean Cyborg 2, the one with the green nucleotide, is the best of the designs? Tell me your oppinion on this. Not sure how to interprent the grey colors of the shapedata.

Here is the
Cyborg 2

Cyborg 2.1
Photo of starryjess

starryjess

  • 35 Posts
  • 2 Reply Likes
My first thought is that Cyborg 2 is better because it's bluer in the neck, so the neck stayed together more strongly. But I wonder if it matters, if the neck stays together anyway, does it matter if it's a little STRONGER? I have also noticed that in Cyborg 2 it's a little bluer in the multiloop than Cyborg 2.1 (where it's supposed to be yellow). Maybe it's a trade off.
Photo of starryjess

starryjess

  • 35 Posts
  • 2 Reply Likes
Well, I looked at it again and Cyborg 2 is only bluer in the multiloop in ONE spot, and Cyborg 2.1 is bluer in the multiloop in other spots. So overall I think Cyborg 2 might be a little bit better. Sorry if I'm being confusing!
Photo of Eli Fisker

Eli Fisker

  • 2239 Posts
  • 498 Reply Likes
Starry, appreaciated your comments!

It is actually great fun, if Cyborg 2 is the best of the two. When it comes to energy, it seems we have some sort of ”ghost nucleotides” in the hookarea. They don't change energy, they don't change score, but they somehow can ad changes in stability. (I think this may be the same for the multiloop ring where nucleotides also can be added, without changing the collective energy.)

I wrote in the beginning of this getsat post:
If you ad anything in the hook area, it mess up your ability to read the design in both Nupack meltplot and RNA fold. It changes the overall appearance of the design in fx. positional entropy (turns overall red, which normally is good, but here tells you nothing) and the curve in meltplot, so you can't see if your design is good or bad.

The more correct conclusion to this should be: If you ad a nucleotide in the hookarea that sometimes makes the entropy look more red in RNAFold or more stable in NUPACK meltcurve, it may just be because that this nucleotide DID make the design a little more stable. (We don't know for sure yet, we need to look more at those neck nucleotides that do change the designs appearance i RNA fold)

The truely funny thing about the two Cyborg designs, is that they are indistinguishable in both in the NUPACK meltcurve and all the RNA fold data.

I sort of assumed that because those ”ghost nucleotides” don't change energy, they don't change anything. I was wrong.
Photo of Eli Fisker

Eli Fisker

  • 2239 Posts
  • 498 Reply Likes
We still don't know how these "ghost nucleotides" work yet, I suspect though that they can be very useful. To add stability to the overall design. It is not always you want to alter the energy level in the hook, just to ad stability. It depend of what sort of neck you are using. A ghost nucleotide might be the solution, to keep the neck you are working with and still ad stability.