The NIGMS Protein Structure Initiative

1
The NIGMS Protein Structure Initiative

• NSF Workshop on the Role of Theory in Biological
  Physics and Materials
• May 16-18, 2004
• Jerry Li
• Program Director
• Division of Cell Biology and Biophysics
• National Institute of General Medical Sciences

2
National Institute of General Medical Sciences
(NIGMS)

• NIH component for basic biomedical research
  support
• FY2004 annual budget 1,905 million
• 80 NIGMS budget is for R01s and P01s
• About 300 million of NIGMS budget supports
  biophysical research
• NIGMS provides half of NIH total research grant
  support for structural biology

3
The Protein Structure Initiative(PSI)

• Inspired by the Human Genome Project
• Enabled by the recent NIH budget doubling
• Faster and cheaper acquisition of structures
  would not be available otherwise
• Aimed to make 3D structural information available
  for most of the identified genes
• Envisioned to be a 10 year project
• Started in 2000 with an initial budget of 30
  million

4
PSI Pilot Research Centers

• Nine research centers 5 year awards for
  2000-2005
• Construction and automation of structural
  genomics pipelines
• Structural determination of unique protein
  structures
• Methodology and technology development
  increases in efficiency and success rates, lower
  costs

5
PSI Pilot Research Centers
p
UK
UK, Japan, Israel
6
Automation in the PSI Centers
Ian A. Wilson, Scripps Research Institute, P50
GM062411
7
(No Transcript)
8
Productivity of the PSI Centers

• Year Structures Cost/Structure
• 01 50 600k
• 02 100 400k
• 03 200 250k
• 04 300-350 180-210k
• estimated

9
Features of the Production phase PSI

• Cooperative Agreement between the NIH and the
  grantees
• Three-component PSI research network
• Large-scale production centers
• Specialized centers for difficult proteins and
  further technology develop
• Disease-oriented centers
• Centralized PSI knowledgebase
• Award in July, 2005

10
Other Programs in Support of the PSI

• Support for investigator initiated technology
  development research aimed at eliminating the
  bottlenecks
• Support for national synchrotron beamline
  construction and upgrade
• Support for functional studies of proteins with
  structures solved by the PSI centers
• New program to promote protein structure modeling

11
Structure Production Goal of PSI-2

• Experimental structure determination for protein
  family representatives
• ( 5,000 proteins)
• Computational structure prediction for the
  homologues
• ( 1,000,000 proteins)

12
General Schemes for Protein Structure Prediction

• Homology to known structure based prediction
  (homology modeling, comparative modeling)
• Physics law based prediction
  (ab initio, de novo)
• Combination of the two
• (threading)

13
Quality of comparative modeling depends
critically on the sequence identity between the
modeling targets and templates
D. Baker A. Sali Science 294, 93, 2001
14
Improvement of Model Quality in the CASP
Competitions
C. Venclovas, J. Moult et al.,Proteins 53, 585,
2003
15
High Accuracy Comparative Modeling Workshop
(October 20-21, 2003)

• Session I Introduction
• Burkhard Rost, Sung-Hou Kim, John
  Moult, Eric Jakobsson
• Session II Proposed Solutions
• David Baker, Michael Levitt,
  Richard Friesner, Barry Honig, Andrej Sali,
  Roland Dunbrack,
• Wilfred van
  Gunsteren
• Session III Uses of Structural Models
• Brian Shoichet, Cathy Peishoff,
  Jeffrey Skolnick, Steven Brenner
• Session IV Membrane Protein Modeling
• Harel Weinstein, Stanley Krystek
• Session V New Ideas Contributions
• Jay Ponder, Matt Jacobson, Bonnie
  Berger,
• Diana Murray
• Session VI Vision Development
• Structured Discussion

16
Workshop Recommendations

• Benchmarks, standards, measures of success
• Alignment Dunbrack
• Refinement Baker
• Force Field Friesner
• Collaborations and new blood
• physicist, statistician, mathematician, computer
  scientist
• 3. Hybrid approaches
• NMR data guided modeling, models used in
  molecular replacement
• Resource needs
• computing power, open source policy, well
  annotated maintained model database

17
Protein Homology Modeling Program

• P20 Exploratory Center mechanism
• Three to four centers with annual budget 750k
• Highly collaborative between traditional modelers
  and theoreticians from other scientific
  disciplines
• Highly innovative projects
• Aim to achieve x-ray crystal structure quality
  for the models
• Plans to establish standards for model quality
  assessment
• Plans to promote utilization of the models