Is the Dot Plot and Melt Point using 37deg or 24deg for the cloud shape labs?

  • 1
  • Question
  • Updated 4 years ago
  • Answered
Is the Dot Plot and Melt point used in game using 37 deg or 24deg for the SHAPE cloud labs?  I was told by Rhiju that the SHAPE labs were done at 24 deg C. RNAfold has 37deg as the default and I read somewhere that setting is being used by Eterna but cant find it now. If it is using 37 deg why? I am going to do an analysis with the temp setting set to 24deg C for RNAfold to see if that shows anything. I was thinking that maybe it was because if something folds at 37 deg you know it will fold at 24 deg. I appreciate the help.
Photo of Jennifer Pearl

Jennifer Pearl

  • 197 Posts
  • 31 Reply Likes

Posted 4 years ago

  • 1
Photo of Jennifer Pearl

Jennifer Pearl

  • 197 Posts
  • 31 Reply Likes
So i cant edit this post so I will clarify that Rhiju said that the SHAPE labs are measured and evaluation at 24deg C.
Photo of nando

nando, Player Developer

  • 388 Posts
  • 71 Reply Likes
The in silico predictions (MFEs and partition functions) are computed for 37°C (which is the baseline used by parameter sets like Turner 1999 and 2004)
The in vitro experiments (I'm sure about the SHAPE mapping experiments, I don't have precise knowledge of how things are done with the microarray pipeline) were done at room temperature (about 25°C).

My guess about the "why" is: convenience. As indicated above, predictions are easier and more naturally made at 37°C. I suppose it is the same reasoning for experiments: it is easier to simply work at room temperature and have one less equipment to monitor (some kind of bath maintained at 37°C). As to whether this makes a big difference or none at all, I don't really know.
Photo of nando

nando, Player Developer

  • 388 Posts
  • 71 Reply Likes
All I can say is: possibly.
My personal opinion is that this most likely isn't the most important factor, but again, I'm not a scientist, so my thoughts on the subject are everything but authoritative. 
Photo of Jennifer Pearl

Jennifer Pearl

  • 197 Posts
  • 31 Reply Likes
I did a quick look at the predicted structures at 24 deg for Simple RNA 
Switch using DPAT. There is a striking different on what is in the predicted Structure. I think it is the MFE (lower triangle) (nando pleas correct me if I am wrong that the MFE is represented by the lower triangle ). At 24 deg there is more leaking of one state into the other state. By leaking I mean that a stack that should form in state 2 is being predicted in state 1 (although a much lower probability). I have called them ghosts in the past. I am going to include this in the paper I am working on. There are many more columns as well and more features I would consider of interest as well
(Edited)
Photo of nando

nando, Player Developer

  • 388 Posts
  • 71 Reply Likes
@jenn: a last note I can contribute about the topic (and the reason why I believe that this difference doesn't matter much): I recall trying once in Cloud Lab to have Vinnie submit designs tuned for 24°C instead of 37°C, and Vinnie has always used some form of ensemble optimization. I saw no improvement at all in the results, if anything, Vinnie scored a little lower than usual that time.
Photo of Jennifer Pearl

Jennifer Pearl

  • 197 Posts
  • 31 Reply Likes
Hi nando,

Not sure if you can answer this but is there a thermodynamic event that is going on during folding that could cause an increase in heat? Maybe that is why at 24 deg it did not work as well? This may be a Brourd question.

Respectfully,
Jennifer
Photo of Brourd

Brourd

  • 437 Posts
  • 79 Reply Likes
The temperature simply represents an average for the system. There will be instances where the kinetic energy of one molecule may be considerably higher than others within the same system, but it doesn't necessarily reflect a condition that would alter the resulting "average" MFE structure of the molecule at that temperature.

Typically, the result of a system with a lower average temperature will be that the particles are not moving quite as much on average. An example may be that a specific nucleotide is not as mobile as its counterpart in a solution of a higher temperature, or tertiary contacts that would form at one temperature are unstable at another temperature.

Theoretically, this difference will result in some interactions possibly being more stable or less stable, given the amount of energy required for the same bond to break would need to be greater or less than that same amount of energy at a different temperature.

While I don't personally know the details as to how the parameters scale between the different temperatures, I would venture a guess and say that each and every possible fold still exists in solution, and that the primary change in prediction is in how helices are predicted to fold or unfold, as the free energies of base pairs and terminal loops are converted.

Granted, the Das Lab's experimental protocol is more complex than a single temperature value. While they may let the RNA fold at room temperature, they first denature the polymer, and then allow it to rest in a water bath that is in the high 40's (if I recall correctly) before finally lowering the temperature to RT. How this alters the thermodynamics of the system has to be taken into account as well.
Photo of Jennifer Pearl

Jennifer Pearl

  • 197 Posts
  • 31 Reply Likes
Thanks Nando