Drawing RNA Secondary Structures

Alexander Batelaan Author

04/05/2021 Added

15 Plays

Description

We improved the drawing of RNA secondary structures by developing algorithms which prevent nucleotide overlap.

Searchable Transcript

Search:

[00:00:02.260]Hello.
[00:00:03.093]My name is Alexander Batelaan
[00:00:04.617]and I will be presenting the research
[00:00:06.910]that I've been working on this school year,
[00:00:08.990]which is drawing RNA secondary structures.
[00:00:15.190]RNA is a very essential molecule in cell biology.
[00:00:18.900]It transports genetic instructions stored in DNA
[00:00:21.730]to the cytoplasm to generate proteins.
[00:00:24.780]RNA can fold into complex structures
[00:00:27.090]which have the potential to be used for nanomedicine.
[00:00:31.100]In our research,
[00:00:32.150]we want to improve the program
[00:00:33.730]that can draw the secondary structure of RNA.
[00:00:37.590]The program called rna_draw draws the secondary structure
[00:00:40.980]of RNA but the structure have nucleotides that overlap.
[00:00:45.230]We worked on finding solutions that prevent overlaps
[00:00:48.160]of nucleotides in the secondary structure of RNA.
[00:00:53.020]It would be helpful to efficiently draw the structure
[00:00:55.890]of RNA because RNA is biologically important.
[00:00:59.570]This image shows how the RNA is part of the process
[00:01:03.000]of decoding genetic information to build proteins.
[00:01:06.850]RNA folding is essential for its biological functions.
[00:01:12.293]RNA can fold according to primary,
[00:01:15.480]secondary, and tertiary structure.
[00:01:18.070]Primary structure is the sequence of the RNA nucleotides.
[00:01:21.740]Secondary structure is how the sequence
[00:01:23.710]forms base pairs in a 2D space,
[00:01:26.710]and tertiary structure is how the 2D structure folds in
[00:01:29.770]on itself to make a 3D shape.
[00:01:32.340]There are programs that can draw RNA secondary structures.
[00:01:35.870]The rna_draw Python files developed
[00:01:38.060]by the Yesselman Research Group is an example
[00:01:40.970]of a program that can draw RNA secondary structures.
[00:01:46.600]I've been working on improving
[00:01:48.110]this secondary structure drawing of RNA molecules.
[00:01:51.810]The rna_draw program currently draws structures
[00:01:54.436]which have nucleotides that overlap.
[00:01:57.730]We have focused on trying to prevent overlaps
[00:02:00.440]in the RNA secondary structure drawings.
[00:02:03.730]We found an algorithm called NEATO,
[00:02:06.210]which solves many of the overlaps.
[00:02:08.920]The NEATO algorithm places an ideal string between each pair
[00:02:13.000]of nucleotides and then uses an iterative solver
[00:02:17.190]to find a low energy configuration.
[00:02:20.700]This configuration often leads
[00:02:22.480]to graphs with reasonable layouts.
[00:02:24.910]In the picture on the right,
[00:02:26.460]we can see a large RNA structure,
[00:02:29.060]the ribosome,
[00:02:30.100]which is drawn with rna_draw
[00:02:32.490]and without the NEATO algorithm.
[00:02:35.700]It looks very messy and has many overlaps.
[00:02:39.290]Then we have the same structure drawn
[00:02:41.410]with rna_draw and the NEATO algorithm.
[00:02:44.840]This figure has much less overlaps
[00:02:47.000]and it looks much better.
[00:02:49.930]But even with the NEATO algorithm,
[00:02:52.410]the secondary structure still has problems.
[00:02:54.950]One of the problems was
[00:02:55.960]that the RNA structures had twists in them.
[00:02:59.090]We had to develop an algorithm
[00:03:00.620]that could detect where these twists occur.
[00:03:04.070]Our algorithm works by looking for line overlaps.
[00:03:07.690]The twist occurs where the lines overlap.
[00:03:10.860]We'll call our algorithm
[00:03:12.540]the overlap detection algorithm.
[00:03:15.210]The overlap detection algorithm looks through all
[00:03:18.070]of the coordinates of the nucleotides.
[00:03:20.540]If we have four points then we can define two lines
[00:03:23.810]and figure out where these lines overlap by using geometry.
[00:03:28.460]Using this algorithm,
[00:03:29.610]we were able to detect all instances where the twists
[00:03:32.750]in the RNA structures occurred.
[00:03:38.470]The next step was to use the information
[00:03:40.684]about where the twists occur in order to undo the twists.
[00:03:44.770]We use the information from the overlap detection algorithm
[00:03:47.920]to define a reflection line going through the twisted
[00:03:50.450]section of the RNA structure.
[00:03:53.080]Then we reflected the part of the RNA structure
[00:03:55.470]that was twisted over the reflection line in order
[00:03:58.380]to undo the twist.
[00:04:00.450]We will call this operation the reflection algorithm.
[00:04:04.890]We used a combination of the overlap detection algorithm
[00:04:07.977]and the reflection algorithm in order
[00:04:10.260]to undo the twists in the RNA structures.
[00:04:13.920]Here are some examples where the algorithms worked.
[00:04:17.070]The first example has one twist which was resolved.
[00:04:20.600]The second example has two twists that were resolved.
[00:04:27.330]In conclusion,
[00:04:28.280]the NEATO algorithm worked well to solve many overlaps
[00:04:31.550]in the drawing of RNA secondary structures
[00:04:35.050]but it still generated structures with twists in them.
[00:04:38.540]So we developed an overlap detection algorithm
[00:04:41.080]which worked well to detect where the twists occurred.
[00:04:44.230]We also developed a reflection algorithm
[00:04:47.070]which reflected parts of the RNA structure
[00:04:49.370]to undo the twists.
[00:04:51.490]The combination of the NEATO algorithm,
[00:04:53.462]overlap detection algorithm,
[00:04:56.200]and reflection algorithm worked
[00:04:58.194]in drawing some example RNA structures
[00:05:01.070]which didn't contain any overlaps.
[00:05:04.320]In the future we will focus
[00:05:06.370]on developing the algorithms further.
[00:05:08.830]The algorithms were not able
[00:05:10.730]to fully work on larger RNA structures.
[00:05:14.800]In the image,
[00:05:15.633]we see that our algorithms made the structures look worse
[00:05:18.509]by putting a kink into one
[00:05:20.190]of the branches of the RNA structure.
[00:05:23.420]We need to look more closely
[00:05:24.780]at what is going on in the RNA structures,
[00:05:28.620]figure out what is going wrong,
[00:05:30.310]and fix the problem.
[00:05:34.410]Here are my references and I would like to thank UCARE
[00:05:37.350]for sponsoring this research in the summer.
[00:05:40.380]I would also like to thank Dr. Yesselman
[00:05:43.050]for allowing me to continue working
[00:05:45.000]on the problem during the school year
[00:05:46.740]for Chemistry 399 research credit.
[00:05:50.910]Lastly, I would like to thank both Dr. Yesselman
[00:05:54.350]and Chris Jurich for mentoring me
[00:05:56.280]and helping me with learning all of the Python programming.
[00:05:59.740]Thank you for listening to my presentation.

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Comments

0 Comments

Drawing RNA Secondary Structures

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments