Protein Structure Prediction 11/11/05 - Iowa State Universityweb.cs.iastate.edu/~cs544/Lectures/1111ProteinStructurePrediction.pdfProtein Structure Prediction 11/11/05 D Dobbs ISU

Protein Structure Prediction 11/11/05

D Dobbs ISU - BCB 444/544X 1

11/11/05 D Dobbs ISU - BCB 444/544X: Protein Structure Prediction 1

11/11/05

Protein StructurePrediction & Modeling


Bioinformatics SeminarsNov 11 Fri 12:10 BCB Seminar in E164 Lago

Building Supertrees Using DistancesSteve Willson, Dept of Mathematicshttp://www.bcb.iastate.edu/courses/BCB691-F2005.html

Next week - Baker Center/BCB Seminars: (seminar abstracts available at above link)

Nov 14 Mon 1:10 PM Doug Brutlag, StanfordDiscovering transcription factor binding sites

Nov 15 Tues 1:10 PM Ilya Vakser, Univ Kansas Modeling protein-protein interactions both seminars will be in Howe Hall Auditorium


Protein Structure & Function:Analysis & Prediction

Mon Protein structure: basics; classification,databases, visualization

Wed Protein structure databases - cont.

Thurs Lab Protein structure databases Visualization software Secondary structure prediction

Fri Protein structure prediction Protein-nucleic acid interactions


Reading Assignment (for Mon-Fri)

Mount Bioinformatics• Chp 10 Protein classification & structure prediction

http://www.bioinformaticsonline.org/ch/ch10/index.html

• pp. 409-491• Ck Errata: http://www.bioinformaticsonline.org/help/errata2.html


BCB 544 Additional Reading

Required:• Gene Prediction:• Burge & Karlin 1997 JMB 268:78

Prediction of complete gene structures in human genomic DNA

Optional:

• Structure Prediction:• Schueler-Furman…Baker 2005 Science 310:638

Progress in modeling of protein structures and interactions


Review last lecture:

Protein Structure:Databases, Classification &

Visualization




Protein sequence databases

• UniProt (SwissProt, PIR, EBI)http://www.pir.uniprot.org

• NCBI Protein http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Protein

More on these later: protein function prediction


Protein sequence & structure: analysis• Diamond STING Millennium - many useful structure

analysis tools, including Protein Dossierhttp://trantor.bioc.columbia.edu/SMS/

• SwissProt (UniProt)protein knowledgebasehttp://us.expasy.org/sprot

• InterPROsequence analysis toolshttp://www.ebi.ac.uk/interpro


Protein structure databases• PDB Protein Data Bank http://www.rcsb.org/pdb/ (RCSB) - THE protein structure database

• MMDB Molecular Modeling Databasehttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Structure

(NCBI Entrez) - has "added" value

• MSD Molecular Structure Database http://www.ebi.ac.uk/msdEspecially good for interactions, binding sites


Protein structure classification• SCOP = Structural Classification of Proteins

Levels reflect both evolutionary and structural relationshipshttp://scop.mrc-lmb.cam.ac.uk/scop

• CATH = Classification by Class, Architecture, Topology & Homologyhttp://cathwww.biochem.ucl.ac.uk/latest/

• DALI/FSSP (recently moved to EBI & reorganized)• fully automated structure alignments

• DALI server http://www.ebi.ac.uk/dali/index.html• DALI Database (fold classification)

http://ekhidna.biocenter.helsinki.fi/dali/start


Protein structure visualization• Molecular Visualization Freeware:

http://www.umass.edu/microbio/rasmol

• MolviZ.Orghttp://www.umass.edu/microbio/chime

• Protein Explorer http://www.umass.edu/microbio/chime/pe/protexpl/frntdoor.htm• RASMOL (& many decendents: Protein Explorer,PyMol, MolMol, etc.)

http://www.umass.edu/microbio/rasmol/index2.htm• CHIME

http://www.umass.edu/microbio/chime/getchime.htm

• Cn3Dhttp://www.biosino.org/mirror/www.ncbi.nlm.nih.gov/Structure/cn3d/

• Deep View = Swiss-PDB Viewerhttp://www.expasy.org/spdbv


Protein structure visualizationSuperb interactive structure visualization software by Jane & Dave Richardson, Duke University

• KINIMAGEhttp://kinemage.biochem.duke.edu/

• Fantastic research tools for structure analysis & refinement




RCSB PDB - Beta sitehttp://pdbbeta.rcsb.org/pdb/Welcome.do


MMDBhttp://www.ncbi.nlm.nih.gov/Structure/MMDB/mmdb.shtml


Cn3Dhttp://www.ncbi.nlm.nih.gov/Structure/CN3D/cn3d.shtml


Cn3D : Displaying 2' Structures


Cn3D: Structural Alignments


SCOP - Structure Classification http://scop.mrc-lmb.cam.ac.uk/scop




6 main classes of protein structure1) α Domains

• Bundles of helices connected by loops

2) β Domains• Mainly antiparallel sheets, usually with 2 sheets forming

sandwich

3) α/β Domains• Mainly parallel sheets with intervening helices, also

mixed sheets

4) α+β Domains• Mainly segregated helices and sheets

5) Multidomain (α & β)• Containing domains from more than one class

6) Membrane & cell-surface proteins


CATH - Structure Classification http://cathwww.biochem.ucl.ac.uk/latest/


Structural Genomics

~ 30,000 "traditional" genes in human genome(not counting: ???)

~ 3,000 proteins in a typical cell> 2 million sequences in UniProt> 33,000 protein structures in the PDB Experimental determination of protein structure

lags far behind sequence determination!Goal: Determine structures of "all" protein folds in nature, using

combination of experimental structure determination methods(X-ray crystallography, NMR, mass spectrometry) & structureprediction


Structural Genomics Projects

TargetDB: database of structural genomics targetshttp://targetdb.pdb.org

Protein Structure Prediction?


Protein Folding

"Major unsolved problem in molecular biology"

In cells: spontaneousassisted by enzymesassisted by chaperones

In vitro: many proteins fold spontaneously & many do not!


Steps in Protein Folding1- "Collapse"- driving force is burial of hydrophobic aa’s

(fast - msecs)2- Molten globule - helices & sheets form, but "loose"

(slow - secs)3- "Final" native folded state - compaction, some 2'

structures rearranged

Native state? - assumed to be lowest free energy - may be an ensemble of structures




Protein Dynamics

• Protein in native state is NOT static• Function of many proteins depends on conformational

changes, sometimes large, sometimes small• Globular proteins are inherently "unstable"

(NOT evolved for maximum stability)• Energy difference between native and denatured

state is very small (5-15 kcal/mol)(this is equivalent to 1 or 2 H-bonds!)

• Folding involves changes in both entropy & enthalpy


Protein Structure Prediction

• Structure is largely determined by sequence BUT:

• Similar sequences can assume different structures• Dissimilar sequences can assume similar structures• Many proteins are multi-functional• Protein folding:

• determination of folding pathways• prediction of tertiary structure

still largely unsolved problems


New today:

Protein Structure Prediction

Secondary structure(text focuses on this - I won't)

Tertiary structure(let's do this instead!)


Deciphering the Protein Folding Code

• Protein Structure Prediction or "Protein Folding" Problem

given the amino acid sequenceof a protein, predict its3-dimensional structure (fold)

• "Inverse Folding" Problemgiven a protein fold, identifyevery amino acid sequencethat can adopt its3-dimensional structure


Protein Structure Determination?High-resolution structure determination

• X-ray crystallography (<1A°)• Nuclear magnetic resonance (NMR) (~1-2.5A°)

Lower-resolution structure determination• Cryo-EM (electron-microscropy) ~10-15A°

Theoretical Models?• Highly variable - now, some equiv to X-ray!


Tertiary Structure PredictionFold or tertiary structure prediction

problem can be formulated as a searchfor minimum energy conformation• search space is defined by psi/phi angles of backbone

and side-chain rotamers• search space is enormous even for small proteins!• number of local minima increases exponentially

of the number of residues

Computationally it is an exceedingly difficult problem!




Ab Initio Prediction1. Develop energy function

• bond energy• bond angle energy• dihedral angle energy• van der Waals energy• electrostatic energy

2. Calculate structure by minimizing energy function(usually Molecular Dynamics or Monte Carlo methods)

Ab initio prediction - not practical in general• Computationally? very expensive• Accuracy? Usually poor for all but short peptides

(but see Baker review!)

Provides both folding pathway & folded structure


Comparative Modeling

Provide folded structure only

Two primary methods1) Homology modeling2) Threading (fold recognition)

Note: both rely on availability of experimentallydetermined structures that are "homologous" orat least structurally very similar to target


Homology Modeling1. Identify homologous protein sequences

• PSI-BLAST• multiple sequence alignment (MSA)

2. Among those with available structures, chooseclosest sequence match for template

3. Build model by placing residues into correspondingpositions of homologous structure models & refineby "tweaking"

Homology modeling - works "well"• Computationally? not very expensive• Accuracy? higher sequence identity ⇒ better model

Requires >30% sequence identity


Threading - Fold RecognitionIdentify “best” fit between target sequence & template structure

1. Develop energy function2. Develop template library3. Align target sequence with each template & score4. Determine best score (1D to 3D alignment)5. Build refine structure as in homology modeling Threading - works "sometimes"

• Computationally? Can be expensive or cheap, depends on energyfunction & whether "all atom" or "backbone only" threading

• Accuracy? in theory, should not depend on sequenceidentity (should depend on quality of template library & "luck")

• But, usually higher sequence identity ⇒ better model


1. Align target sequence with template structures(fold library) from the Protein Data Bank (PDB)

2. Calculate energy (score) to evaluate goodness of fitbetween target sequence & template structure

3. Rank models based on energy scores

TargetSequence

StructureTemplates

ALKKGF…HFDTSE

Threading - a "local" example


Threading Goals & Issues

Structure database - must be complete: no decent model if nogood template in library!

Sequence-structure alignment algorithm:Bad alignment ⇒ Bad score!

Energy function (scoring scheme):• must distinguish correct sequence-fold alignment from

incorrect sequence-fold alignments• must distinguish “correct” fold from close decoys

Prediction reliability assessment - how determine whetherpredicted structure is correct (or even close?)

Find “correct” sequence-structure alignment of atarget sequence with its native-like fold in PDB




Threading – Structure database

Build a template database (e.g., ASTRAL domain library derived from PDB)

Supplement with additional decoys, e.g., generated usingab initio approach such as Rosetta (Baker)


Threading - Energy function

Two main methods (and combinations of these)

• Structural profile (environmental) physico-chemical properties of aa’s

• Contact potential (statistical)based on contact statistics from PDB

(Miyazawa & Jernigan - Jernigan now at ISU)


Protein Threading – typical energy functionMTYKLILNGKTKGETTTEAVDAATAEKVFQYANDNGVDGEWTYTE

How well does aspecific residue fitstructural environment?

What is "probability"that two specificresidues are incontact?

Alignment gappenalty?

Total energy: E_p + E_s + E_g

Find a sequence-structure alignment that minimizingthe energy function


A Rapid Threading Approach forProtein Structure Prediction

Kai-Ming Ho, PhysicsHaibo Cao Yungok Ihm

Zhong GaoJames MorrisCai-zhuang Wang

Drena Dobbs, GDCBJae-Hyung LeeMichael TerribiliniJeff Sander


Performance Evaluation? "Blind Test"

CASP5 Competition (Critical Assessment of Protein Structure Prediction)

Given: Amino acid sequence Goal: Predict 3-D structure (before experimental results published)

Typical Results: well, actually, BEST Results:

HO = #1 ranked CASP prediction for this target

Target 174PDB ID = 1MG7

Actual Structure

Predicted Structure

T174_1

T174_2



FR Fold Recognition(targets manually assessed by Nick Grishin)

-----------------------------------------------------------

Rank Z-Score Ngood Npred NgNW NpNW Group-name 1 24.26 9.00 12.00 9 12 Ginalski 2 21.64 7.00 12.00 7 12 Skolnick Kolinski 3 19.55 8.00 12.50 9 14 Baker 4 16.88 6.00 10.00 6 10 BIOINFO.PL 5 15.25 7.00 7.00 7 7 Shortle 6 14.56 6.50 11.50 7 13 BAKER-ROBETTA 7 13.49 4.00 11.00 4 11 Brooks 8 11.34 3.00 6.00 3 6 Ho-Kai-Ming 9 10.45 3.00 5.50 3 6 Jones-NewFold -----------------------------------------------------------

FR NgNW - number of good predictions without weighting for multiple modelsFR NpNW - number of total predictions without weighting for multiple models

Overall Performance in CASP5 Contest (M. Levitt, Stanford)

Documents

Protein Structure Prediction 11/11/05 - Iowa State Universityweb.cs.iastate.edu/~cs544/Lectures/1111ProteinStructurePrediction.pdfProtein Structure Prediction 11/11/05 D Dobbs ISU