Problem Sequence Research

  • 2
  • Article
  • Updated 8 years ago
is a potential problem sequence. I noticed it first in this design and kept an eye out; it came up again as a problem in the beautiful 99/100 design. So I set up a spreadsheet to find what (I think) are all the examples of places where the problem sequence has been actually synthesized by the eteRNA lab for Lab 103.

Any assistance in looking at all of them and seeing whether they seemed to cause problems would be great!

Here they are. The numbers show some base numbers that the sequence starts at.
11,80,10,81 and 62,71,61,72 [link]
42,49,41,50 [link]
42,49,41,50 [link]
42,49,41,50 [link]
60,73,59,74 [link]
42,49,41,50 [link]
42,49,41,50 [link]
11,80,10,81 and 62,71,61,72 [link]
42,49,41,50 [link]
42,49,41,50 [link]
42,49,41,50 [link]
42,49,41,50 [link]
42,49,41,50 [link]
11,80,10,81 and 62,71,61,72 [link]
42,49,41,50 [link]
10,81,9,82 [link]
33,16,32,17 [link]
8,83,7,84 [link]
60,73,59,74 [link]
32,17,31,18 [link]
32,17,31,18 [link]
42,49,41,50 [link]
Photo of Chris Cunningham [ccccc]

Chris Cunningham [ccccc]

  • 97 Posts
  • 13 Reply Likes
  • scientific?

Posted 8 years ago

  • 2
Photo of Ding

Ding

  • 94 Posts
  • 20 Reply Likes
Hi ccccc

I'd be happy to take on a few of these. I think it might make sense to take them on in batches of related designs (which can pretty much be determined from the positions of the CAU sequence as listed above).

Okay if I start with Fabian's three designs (with CAU at 9,10,11 and 60, 61, 62; #1, #8, and #14 on your list)?

EDIT TO ADD:
Okay, I've taken a preliminary look at the three designs by Fabian where these triplets each occur twice ("1337", "Even More 1337", and "1337 - Improved") and they don't seem to cause any problems.

In all occurrences, all six nucleotides involved are bonded, and in only one instance is there an unbonded pair next to this sequence (in "Even More 1337" the U at 74 and A at 75, though the nucleotides they're paired with - A at 59 and U at 58 - measure as bonded).
Photo of Chris Cunningham [ccccc]

Chris Cunningham [ccccc]

  • 97 Posts
  • 13 Reply Likes
Agh -- I can't edit the original post? That's absurd. I'll track progress in this post.

Has 0/6 unbonded, checked by Ding 11,80,10,81 and 62,71,61,72 [link]
Has 1/6 unbonded, checked by ccccc 42,49,41,50 [link]
Has 0/6 unbonded, checked by ccccc 42,49,41,50 [link]
Has 1/6 unbonded, checked by ccccc 42,49,41,50 [link]
Has 2/6 unbonded, checked by ccccc 60,73,59,74 [link]
Has 2/6 unbonded, checked by chaendryn 42,49,41,50 [link]
Has 1/6 unbonded, checked by chaendryn 42,49,41,50 [link]
Has 0/6 unbonded, checked by Ding 11,80,10,81 and 62,71,61,72 [link]
Has 0/6 unbonded, checked by chaendryn 42,49,41,50 [link]
Has 0/6 unbonded, checked by chaendryn 42,49,41,50 [link]
Has 0/6 unbonded, checked by chaendryn 42,49,41,50 [link]
Has 0/6 unbonded, checked by chaendryn 42,49,41,50 [link]
Has 1/6 unbonded, checked by chaendryn 42,49,41,50 [link]
Has 0/6 unbonded, checked by Ding 11,80,10,81 and 62,71,61,72 [link]
Has 0/6 unbonded, checked by chaendryn 42,49,41,50 [link]
Unreadable, checked by ccccc 10,81,9,82 [link]
Has 0/6 unbonded, checked by chaendryn 33,16,32,17 [link]
No shape data for 2 of the bases 8,83,7,84 [link]
Has 0/6 unbonded, checked by chaendryn 60,73,59,74 [link]
Has 0/6 unbonded, checked by chaendryn 32,17,31,18 [link]
Has 2/6 unbonded, checked by chaendryn 32,17,31,18 [link]
Has 1/6 unbonded, checked by chaendryn 42,49,41,50 [link]
Photo of chaendryn

chaendryn

  • 29 Posts
  • 1 Reply Like
Heya 5c. I'll start working my way up the list from the bottom :) Not sure if I'll be reporting this correctly, but here goes:

# 22 - 26735 (last one on the list) by Donald. Looks like it's a 1/6 unbonded in the CAU sequence starting at 41. The U at 50 is shown as unbonded but it does not appear to distort the shape of the solution.

# 21 - 26634 by singinst - 2/6 unbonded - the UA pair at 32/17 does not bond at all and creates a second bulge in that arm.

# 20 - 26628 by chesterfield - 0/6 unbonded - shape appears stable at that point. Same position as sequence in 26634 - this one was fine, while 26634 was not.

# 19 - 80753 by ccccc - 0/6 unbonded - shape appears stable at that point.

Hope I'm doing this right :) Will check some more later.

Chaen
Photo of chaendryn

chaendryn

  • 29 Posts
  • 1 Reply Like
Okays ... here's some more.

#17 - 214920 by AnticNoise - 0/6 unbonded, shape appears stable at that point.

#15 - 176686 by Nadine - 0/6 unbonded, shape appears stable at that point.

#13 - 293863 by dimension9 - 1/6 unbonded, U at 50, does not appear to distort the shape

#12 - 291417 by mat747 - 0/6 unbonded, shape appears stable at that point.

#11 - 291042 by madde - 0/6 unbonded, shape appears stable at that point.

#10 - 290344 by Berex NZ - 0/6 unbonded, shape appears stable at that point.e

#9 - 278220 by Ding - 0/6 unbonded, shape appears stable at that point.

#7 - 139776 by dimension 9 - 1/6 unbonded, A at 41, does not appear to distort the shape

#6 - 136472 by mat747 - 2/6 unbonded, AU pair at 41/50, creates a second bulge in that arm

That should be all of the missing ones ...

Hope this helps. Though if I can help/improve on it, let me know :)

Chaen
Photo of Chris Cunningham [ccccc]

Chris Cunningham [ccccc]

  • 97 Posts
  • 13 Reply Likes
Excellent! So it looks like we've synthesized this sequence 22 times, and 20 of those had readable relevant shape data. Of those 20 times, it had:

  • 60%: perfect shape [0/6 unbonded]
  • 25%: 1/6 unbonded
  • 15%: 2/6 unbonded


I feel like this is a pretty bad record since all of the problems it causes are local problems (not caused by far away interactions). However it is not immediately clear to me how to decide whether to make a claim that we've learned something or not.

Is this a worse record than an average sequence of this length?

If anyone has ideas on how to proceed please post them; I'll be thinking about it as well.
Photo of Ding

Ding

  • 94 Posts
  • 20 Reply Likes
I can see a couple ways to go forward.

First is to ask why this combination succeeds sometimes and fails others. My first hypothesis based on quickly scanning through all the designs listed would be to check the effect tetraloop stabilization has. I don't think I saw any instances of failure in designs where the tetraloop at the end of the branch this pattern was in was stabilized as xGAAAx or CUUCGG or one of the other known stable tetraloops.

One example is to compare Maelstrom's One Cross Bulge First Attempt to ccccc's Maelstrom's Tempest, where the only difference seems to be tetraloop stabilization and this pattern goes from failing in 2/6 places to 0/6.

The other direction to go is to question whether this pattern is actually any more problematic than other instances where UA AU or AU UA is flanked by a GC or CG pair. One that seemed to pop up a lot as a problem is

CG
UA
AU

Which might be a good place to start as a comparison to

CG
AU
UA
Photo of chaendryn

chaendryn

  • 29 Posts
  • 1 Reply Like
Something that's been rattling around in my head this week ...

What does the data show, if anything, about the similar sequence read from the 3' end rather than the 5' end? (I know this is probably not the correct way to phrase it, but couldn't think of another way to put it.) So we're looking at

GC
UA
AU

and

AU
UA
GC

as possible areas of exploration.

We know that swapping a base pair around has an influence on free energy, so will the sequences above actually prove to be more or less stable in synthesis?

Will go and have a look at the data over the weekend and see if I'm off on a wild goose chase :)

Chaen