Difference between revisions of "Lecture 11"

Latest revision as of 15:27, 1 September 2007

Update Warning! This page has not been revised yet for the 2007 Fall term. Some of the slides may be reused, but please consider the page as a whole out of date as long as this warning appears here.

(Previous lecture) ... (Next lecture)

Protein Structure Prediction

...

Add:

Summary points
Exercises
Further reading

Lecture Slides

Slide 001

Lecture 11, Slide 001

Slide 002

Lecture 11, Slide 002

Slide 003

Lecture 11, Slide 003
The observation became famously known as "Levinthal's Paradox", that neither random search nor the postulation of folding pathways can explain how a discrete structure can self-organize, given in a combinatorially large search space. But real proteins do just fine, thank you.

Slide 004

Lecture 11, Slide 004
A toy observation on an optimization problem of similar magnitude as the random folding of a protein...

Slide 005

Lecture 11, Slide 005

Slide 006

Lecture 11, Slide 006

Slide 007

Lecture 11, Slide 007

Slide 008

Lecture 11, Slide 008

Slide 009

Lecture 11, Slide 009

Slide 010

Lecture 11, Slide 010

Slide 011

Lecture 11, Slide 011

Slide 012

Lecture 11, Slide 012

Slide 013

Lecture 11, Slide 013

Slide 014

Lecture 11, Slide 014

Slide 015

Lecture 11, Slide 015

Slide 016

Lecture 11, Slide 016

Slide 017

Lecture 11, Slide 017

Slide 018

Lecture 11, Slide 018
Non-polynomial time-complexity problems are considered intractable, since even as the problem size 'n' grows only modestly, the time requirements grow beyond all bounds and reasonable resources. A 1,000 element problem of O(2ⁿ) complexity takes the age of the universe to compute.

Slide 019

Lecture 11, Slide 019

Slide 020

Lecture 11, Slide 020

Slide 021

Lecture 11, Slide 021

Slide 022

Lecture 11, Slide 022

Slide 023

Lecture 11, Slide 023

Slide 024

Lecture 11, Slide 024
Simulated annealing allows a system to be computationally moved out of situations where it is trapped in local minima, and to proceed towards a global minimum on a rough search landscape.

Slide 025

Lecture 11, Slide 025
A simulated annealing strategy (in pseudocode).

Slide 026

Lecture 11, Slide 026
Many believe that genetic algorithms - albeit interesting - rarely outperform simulated annealing.

Slide 027

Lecture 11, Slide 027
Natural proteins of course have evolved under the constraint of foldability. They may have avoided mutations that would expose them to the requirements of full, combinatorial optimization of their 3-D structure.

Slide 028

Lecture 11, Slide 028

Slide 029

Lecture 11, Slide 029

Slide 030

Lecture 11, Slide 030

Slide 031

Lecture 11, Slide 031

Slide 032

Lecture 11, Slide 032

Slide 033

Lecture 11, Slide 033

Slide 034

Lecture 11, Slide 034

Slide 035

Lecture 11, Slide 035

Slide 036

Lecture 11, Slide 036

Slide 037

Lecture 11, Slide 037

Slide 038

Lecture 11, Slide 038

@@ Line 1: / Line 1: @@
+<div style="padding: 5px; background: #FF4560;  border:solid 2px #000000;">
+'''Update Warning!'''
+This page has not been revised yet for the 2007 Fall term. Some of the slides may be reused, but please consider the page as a whole out of date as long as this warning appears here.
+</div>
+&nbsp;
+&nbsp;
 __NOTOC__
 <small>[[Lecture_10|(Previous lecture)]] ... [[Lecture_12|(Next lecture)]]</small>
@@ Line 127: / Line 136: @@
 ======Slide 027======
 [[Image:L11_s027.jpg|frame|none|Lecture 11, Slide 027<br>
+Natural proteins of course have evolved under the constraint of foldability. They may have avoided mutations that would expose them to the requirements of full, combinatorial optimization of their 3-D structure.
+]]
-]]
 ======Slide 028======
 [[Image:L11_s028.jpg|frame|none|Lecture 11, Slide 028<br>