A question about the raw SHAPE data

  • 1
  • Question
  • Updated 6 years ago
  • Answered
So I'm gathering all of the SHAPE data at the moment, and I'm wondering if anyone knows about the following.
eg. Lab 3376347
Red Rover III GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGCACACACUUCGGUGUGUGAAAAGAAACAACAACAACAAC
1 2 3 4
-16.8559 -13.5947 -11.3909 -9.8019
222643.8912 222643.4126 222643.1473 222642.985

I'm curious about what that last line represents.
The line with 1,2,3,4 I'm assuming represents the nt position.
The line underneath that represents the SHAPE number.
But what lies under that?

Anyone?
Thanks
Photo of Berex NZ

Berex NZ

  • 116 Posts
  • 20 Reply Likes

Posted 6 years ago

  • 1
Photo of Brourd

Brourd

  • 454 Posts
  • 83 Reply Likes
The third line is the reactivity error for a nucleotide. REACTIVITY Error is defined on the RMDB as

"If REACTIVITY was derived as a consensus of different replicates, this indicates the standard error between samples used. If only one experiment was done, this may be some measure of variation between the data in that experiment, e.g. 0.2 of the standard deviation."
Photo of Berex NZ

Berex NZ

  • 116 Posts
  • 20 Reply Likes
Does that mean, that the SHAPE data for that specific example is unreliable?
Photo of rhiju

rhiju, Researcher

  • 403 Posts
  • 123 Reply Likes
Yes that definitely means unreliable! We find that sequences with lots of G's are very difficult to synthesize and amplify experimentally. In this case, I bet we got no data at all for the sequence...
Photo of Brourd

Brourd

  • 454 Posts
  • 83 Reply Likes
So, Rhiju explains how the error value for the final SHAPE data is determined here.

https://getsatisfaction.com/eternagam...

If I recall correctly, Rhiju mentioned something about 0.8 being the upper bound for reactivity error, where results start to become questionable.

So, 222000 may be a bit high to consider the data reliable :)
Photo of Hyphema

Hyphema

  • 91 Posts
  • 25 Reply Likes
Wow!! That is one high error rate. You figure after about a hundred thousand or so you should just get an error reading of "sorry, back to the drawing board "