Andrea Howells - Summer 2011 Improved Roadmap Sampling Methods for - - PowerPoint PPT Presentation
Andrea Howells - Summer 2011 Improved Roadmap Sampling Methods for High Dimensional Robots (Protein Emulators) Using Molecular Energy Projections The Team Lydia Tapia Assistant Professor Latha Doddikadi - graduate student John
needed to current probabilistic roadmap method software developed for robotic motion planning to map protein folding pathways.
robotic motion planning
approximate map of the molecule's energy landscape
structure formation order, subtle folding differences and relative folding rates can be captured
motion in determining binding affinities between peptide, MHC (Major histocompatability complex), and TCR (T cell receptors) molecules
the specificity of immune responses which can then be used in vaccine research AIDS virus
Caption: This model of the his model of HIV is the most detailed 3-D model of the virus ever made. It summarizes the results from scientific publications in the fields of virology, X-ray analysis and NMR spectroscopy. The depicted spatial configurations of proteins found in HIV particles are in strict accordance with their known 3-D structures. This image relates to the 2010 International Science and Engineering Visualization Challenge appearing in the Feb. 18, 2011, issue of Science, published by AAAS. Credit: Image courtesy of Ivan Konstantinov, Yury Stefanov, Aleksander Kovalevsky, Yegor Voronin - Visual Science Company Usage Restrictions: None
program changing parameters to learn which parameters affect node generation
BasicPRMStategy for pathfinding
using Vizmo, to view the nodes and edges generated
best performance versus cost, using MCH, TCR and 1QLP(1 Alpha Antitrypsin) 1QLP was chosen as a representative protein as it is much than MCH and TCR
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10 5000 10000 15000 20000 25000 30000
Q Chain Angles 0.5 to 10
Pflex: 0.06 Prigid: 0.05 51:48:36 Pflex: 0.06 Prigid: 0.005 38:43:50 Pflex: 0.06 Prigid: 0.0005 43:57:12 Pflex: 0.01 Prigid: 0.0005 78:23:55
Angle Deviations Rejects
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10 5000 10000 15000 20000 25000 30000
Q Chain Angles 0.5 to 10
Pflex: 0.06 Prigid: 0.05 51:48:36 Pflex: 0.06 Prigid: 0.005 38:43:50 Pflex: 0.06 Prigid: 0.0005 43:57:12 Pflex: 0.01 Prigid: 0.0005 78:23:55
Standard Angles Rejects
00:00:00 12:00:00 24:00:00 36:00:00 48:00:00 60:00:00 72:00:00 84:00:00
1QLP Run Times
Pflex: 0.06 Prigid: 0.05 Pflex: 0.06 Prigid: 0.005 Pflex: 0.06 Prigid: 0.0005 Pflex: 0.01 Prigid: 0.0005 0 1 2 3 4 5 6 7 8 9 200 400 600 800 1000 1200 1400 1600 200 400 600 800 1000 1200 1400 1600
Layer Layer size
0 1 2 3 4 5 6 7 8 9 200 400 600 800 1000 1200 1400 1600 200 400 600 800 1000 1200 1400 1600
Layer Layer size
0 1 2 3 4 5 6 7 8 9 200 400 600 800 1000 1200 1400 1600 200 400 600 800 1000 1200 1400 1600
Layer Layer size
0 1 2 3 4 5 6 7 8 9 200 400 600 800 1000 1200 1400 1600 200 400 600 800 1000 1200 1400 1600
Layer Layer size
Free or un-bound multiple robot bodies with articulated links will be used to simulate the protein molecule's energy landscape.
My participation in this project is supported by a DREU (distributed research experiences for undergraduates) internship funded by CRA-W.