Bioinformatics and Computational Molecular Biology Geoff Barton

Bioinformatics and Computational Molecular Biology

Geoff Barton

http://www.compbio.dundee.ac.uk

Practical Tutorial

• Dr David Martin practical tutorial on the use of pymol molecular graphics software.

• In this lecture I will show lots of protein structures – use www.ebi.ac.uk/msd to find them, and/or scop domains database (find with google).

http://www.ebi.ac.uk/msd

Similarities in Proteins

• Lecture 1– Overview of data in molecular biology– Protein modelling– Similarities of Protein Sequence, Structure,

Function

Introduction to Sequence Comparison

• Lecture 2:– Why compare sequences?– Methods for sequence comparison/alignment.– Multiple alignment– Database searching - FASTA/BLAST– Iterative searching - PSI-BLAST

Practical/WWW references

• Organised by Drs Martin– Good preparation would be to look at:

http://www.ebi.ac.uk/Tools andhttp://www.ncbi.nlm.nih.gov

– Look at BLAST and FASTA on these sites as well as database access facilities.

Private DataPast Experiments.Lab note books.

Group discussions.

Traditional biological research

AnalysisReading. Talking.

Thinking.

Hypothesis!

ExperimentDesign. Execution.

Publish!

Public DataJournals

Conferences

Private DataPast Experiments.Lab note books.

Group discussions.DNA sequences

Protein SequencesGenetic mapsTranscripts

3D structuresproteomics results

SNP dataetcetcetc

Bioinformatics/Computational Biology and biological research Analysis

Reading. Talking.Thinking.

ComputationalAnalysis

Software Development

Hypothesis!Computer aided.

ExperimentDesign. Execution.

Computational experimentsSimulation

Publish!Database submission

Database management

Public DataJournals

ConferencesDNA sequences

Protein SequencesGenetic mapsTranscripts

3D structuresproteomics results

SNP dataetcetcetc

EMBL Nucleotide Sequence Database Growth (to 2nd Oct 2006)

Taken from: www.ebi.ac.uk

Protein Sequences

Approx 3,500,000 known for all species (Oct. 2006.)

25,000 for Human

(not counting splice variants and post-translational modifications)

Protein 3D Structures

Approx 39,000 known(much duplication)

Biological data in context

DNA

RNA

Protein Sequence

Protein 3D structure

Molecular function

Overview of Biological Hierarchy...

Whole organismanimal, plant, etc.

Tissue/organbrain, heart, lungs

blood, ...

Ecosystemmany different organisms

Populationgroup of the same type of organism

Familygroup with known common lineage

Cellnerve,muscle,etc..

Organellenucleus, mitochondria, etc...

Nucleus

Chromosome

Gene

MolecularLevels

DNA

RNA

Protein Sequence


Molecular function



blood, ...






Nucleus

Chromosome

Gene

Expression Data(Transcriptomics)

Which of the genes are switched on in which cells/tissues and when?

What are the effects of drugs anddisease on expression patterns

DNA ‘CHIP’ TECHNOLOGY

Technology and data in biology

DNA

RNA

Protein Sequence


Molecular function



blood, ...






Nucleus

Chromosome

Gene

Protein Expression Data(Proteomics)

Which proteins arebeing produced in which cells/tissues when? Which modified forms are present?

What are the effects of drugs and disease on these patterns

2D Gels + Mass Spectrometry.


DNA

RNA

Protein Sequence


Molecular function



blood, ...






Nucleus

Chromosome

Gene

Protein 3D Structure - the bridge to chemistry(Structural Genomics)

What is the atomic level structure of the protein?

What other molecules does it interact with?

What small molecules - potential drugs - does it interact with?

What are the effects of point mutations on the structure?

X-ray crystallography, NMR spectroscopy, single particle, cryo-electron microscopy.




blood, ...






Nucleus

Chromosome

Gene

DNA

RNA

Protein Sequence


Molecular function

Overview of Biological Hierarchy...

Macroscopic Levels

Biology is now a data intensive science

To do good science, you need to know how to use (and not abuse)

computational tools.

Protein Structure Prediction

• ‘Homology’ modelling– Relies on the fact that similarity of sequence

implies similarity of 3D structure.

Lysozyme (1lz1) -lactalbumin (1alc)

?

Imagine we don’t know the 3D structure of -lactalbumin, but we do know its amino acid sequence and that of lysozyme


37.7% Identity, Z=17.6

?

Protein structure prediction(Homology Modelling)

• Align sequence of protein of unknown structure to sequence of protein of known structure.

• In ‘conserved core’ of protein, substitute the amino acid types into the known structure.

• Deal with ‘loops’ between the core elements of structure.



Protein structure prediction(Homology modelling)

• Problems:– Need protein of known structure that is similar

in sequence.– Building loops where there are deletions.– Verifying model.

• Key is getting a good alignment in the first place– Bad alignment => bad model.

Good alignment on its own can:

• Identify key residues (absolutely conserved)

• Identify likely protein core (conserved hydrophobic residues)

• Help predict protein secondary structure (not this lecture).

Sequence alignment is a fundamental technique in

molecular biology.

• May predict proteins of common function even when no 3D structure is known.

• May be used to predict 3D structure and so help understanding of mutants.

• Some examples of where this is right and wrong...

Prediction of structure and function by similarity to known

sequences and structures

Assumption is that similar sequence implies similar structureand function.

But what do we mean by “similar”?

Does similarity of sequence really imply similarity of function?

Protein Sequence/Structure/Function Network

Sequence 3D Structure Function

Similar Similar Similar

Different Different Different





Similar Sequence, Similar Structure, Similar Function.

e.g. Trypsin-like Serine Proteinases

Same fold, same catalytic mechanism.

But DIFFERENT specificity.

e.g. Immunoglobulin variable domains.

Same fold, similar binding function.

But DIFFERENT specificity.

True of all examples. Similarities only give clues to function, differences in specificity can be regarded as differences of function.

ImmunoglobulinVariable Domains

e.g. see: 1a2y

Tryptophan at core of Ig variable domain







-crystallin/L-Lactate Dehydrogenase





Trypsin (3ptn) Subtilisin (2sec)

Trypsin (3ptn) Subtilisin (2sec)

Trypsin (3ptn)

Subtilisin (2sec)

His- 57, Asp-102, Ser-195

Asp- 32, His- 64, Ser-221





Nature 398,84-90, 1999

PDB: 1b47

11% sequence ID

rmsd 1.47Åover 70 residues

PDB: 1b47





Russell, R. B. and Barton, G. J. (1993), "An SH2-SH3 Domain hybrid", Nature, 364, 765.

PDB: 1bia PDB: 2ptk

PDB:2aai PDB:1bas

Matthews, S., et al. (1994), "The p17 Matrix Protein from HIV-1 is Structurally Similar to Interferon-gamma", Nature, 370, 666-668.





Does this ever happen?

HIV Reverse Transcriptase (RT)

HIV Reverse Transcriptase (RT)

HIV Reverse Transcriptase (RT) - domain linkers

Protein Sequence and Structural Similarity

Type Similarity Find By

Homologous(scop family)

Similar StructureSimilar SequenceSimilar Function

Pair-wise SequenceComparison(BLAST/FASTA/Smith-Waterman)

‘RemoteHomologue’(scop superfamily)

Similar StructureWeakly Similar SequenceSimilar Function

ProfileIterative Search(e.g. PSI-BLAST)

Threading/fold recognition?

Analogue(scop fold)

Similar StructureNo sequence similarityOften no functionalsimilarity

Solve BOTH structures byX-ray/NMR methods.

Mapping?










Analogue(scop fold)



Mapping?

Barton, G. J. et al, (1992), "Human Platelet Derived Endothelial Cell Growth Factor is Homologous to E.coli Thymidine Phosphorylase", Prot. Sci., 1, 688-690.










Analogue(scop fold)



Mapping?

Barton, G. J., Cohen, P. T. C. and Barford, D. (1994),"Conservation Analysis and Structure Prediction of the Protein Serine/Threonine Phosphatases: Sequence Similarity with Diadenosine Tetra-phosphatase fromE. coli Suggests Homology to the Protein Phosphatases", Eur. J. Biochem.,220, 225-237.










Analogue(scop fold)



Mapping?


Reading material for this lecture:

This lecture itself. pdf’s for “Barton” papers: www.compbio.dundee.ac.uk/ftp/pdf/

Database statistics: http://www.ebi.ac.uk/embl/

Structure of the amino-terminal domain of Cbl complexed to its binding site on ZAP-70 kinase Wuyi Meng, Sansana Sawasdikosol, Steven J. Burakoff, Michael J. EckNature 398, 84 - 90 (04 March 1999)(available on-line at www.nature.com - search for ZAP-70 kinase - republished in December on-line)

Protein recognition: An SH2 domain in disguise John Kuriyan, James E. DarnellNature 398, 22 - 25 (04 March 1999) (news and views article for above paper)


Matthews, S., et al. (1994), "The p17 Matrix Protein from HIV-1 is Structurally Similar to Interferon-gamma", Nature, 370, 666-668.

Barton, G. J., Cohen, P. T. C. and Barford, D. (1994),"Conservation Analysis and Structure Prediction of the Protein Serine/Threonine Phosphatases: Sequence Similarity with Diadenosine Tetra-phosphatase fromE. coli Suggests Homology to the Protein Phosphatases", Eur. J. Biochem.,220, 225-237.

The end of Lecture 1

Lecture 2 will be on sequence comparison methods.

Documents

Bioinformatics and Computational Molecular Biology Geoff Barton