Introduction to Bioinformatics and its Applications

1
Introduction to Bioinformatics and its
Applications

• Mohamad Rabbath
• Msc. Inf. Freiburg University Software Engineer
  in OFFIS

2
Outline

• Introduction
• State of the Art in Bioinformatics
• Bioinformatics Applications
• Sequences Alignment Phylogenetic Trees
• RNA Algorithms
• Protein Structure Prediction Studies
• Summary
• References
• Demo (Optional), http//cpsp.informatik.uni-freibu
  rg.de8080/index.jsp

3
Introduction

• What is Bioinformatics?

Biological Data
Computing Power

4
Intorduction

• Why do we need computers to analyze biological
  data?
• A very large amount of biological data needs fast
  algorithms and computational resources
• Exponential growth of biological data

5
Intorduction
http//www.ncbi.nlm.nih.gov/Genbank/genbankstats.h
tml
6
Intorduction

• Why do we need computers to analyze biological
  data?
• A Very large ammount of biological data needs
  fast algorithms and computational resources
• Exponential growth of biological data
• Growing gap between known sequences and known
  structures (RNA and Proteins)
• Structure prediction means lower cost

7
Intorduction

• What skills are required for Bioinformatics?
• Deep knowledge in Database design and
  implementation
• Extensive ability of programming and designing
  complex software systems and algorithms (
  languages like c, perl, python, java, are
  widely used in bioinformatics)
• Strong ability of mathematical knowledge and
  statistics
• Some background in biology

8
State of the Art in Bioinformatics

• Bioinformatics was the key element in completing
  the Genome project in 2003
• Currently many researches study the RNA and
  Protein structures
• Sequencing human genome (23andme project is a
  very good example)

9
Bioinformatics Applications

• Drugs discovery and design
• Sequences Alignment
• RNA secondary structure prediction
• Proteins structures prediction
• Reduction of the cost of the Healthcare System
  (Ex. Early discovery of genetic diseases)

10
Sequences Alignment

• Arranging regions of similarity that may be a
  consequence of functionality
• Evolutionary relationships
• Very similar to Natural language processing
  alignment problems

11
Sequences Alignment

• A T G A A C C G C C T A A G C G G C A -- G
• A T G -- -- C C G A C T A -- A C G G A A G
• Three operations are defined
• Substitution
• Insertion
• Deletion

12
Sequences Alignment Needleman-wunsch

• Three steps in dynamic programming
• Initialization
• Matrix fill (scoring)
• Traceback (alignment)
• Mi,j MINIMUM Mi-1, j-1 Si,j (substitution),
  Mi,j-1 w (deletion),
• Mi-1,j w (insertion)

13
Sequences Alignment Needleman-wunsch
14
RNA Secondary Structure Prediction

• RNA plays main rule in Biological Information Flow

15
RNA Secondary Structure Prediction

• RNA GGGCGUGGGCGUAGUCGU
• RNA Structure

16
RNA Secondary Structure Prediction

• C G
• A U
• CAGUCCGGCUGC ..

17
RNA Secondary Structure PredictionNussinov

• Idea Maximizing the number of base pairs

18
RNA Secondary Structure PredictionNussinov
G C A C G A
C G
G
C
A
C
G
A
C
G
19
RNA Secondary Structure PredictionNussinov
G C A C G A
C G
G
C
A
C
G
A
C
G
G C A C G A C G
20
RNA Secondary Structure Prediction Limitation of
Nussinov

• Base pair maximization does not yield
  biologically relevant structures
• Only one structure predicted
• Crossing structures can not be predicted

21
RNA Secondary Structure Prediction Zucker

• Idea Energy minimization

22
Protein Structure Prediction

• The primary structure is a sequence of amino
  acids
• 3D Structure function

23
Protein Structure Prediction HP
(Hydrophobic-Polar) Model

• The chemical group R (the side chain of the amino
  acid) gives the unique properties
• The hydrophobic amino acids tend to cluster
  together
• HP model restricts the 20 amino acids to two
  classes

24
Protein Structure Prediction3D Lattices
Representation
Cubic Lattice
FCC Lattice
25
Protein Structure PredictionHow to achieve
prediction?

• Folding Simulation
• Hidden Markov Model and other stochastic models?
• Statistical model does not guarantee optimality
• Polynomial time required
• Constraint Programming Approach
• Optimality is guaranteed
• NP Completeness makes it time consuming

26
Protein Structure PredictionWhy Why constraint
approach?

• The Solution is not unique (optimality is
  targeted)
• The Space of the problem is relatively small
• Offline problem

27
Protein Structure Prediction Enhanced constraint
approach
28
Protein Structure PredictionWhy Degenerecy and
protein-like sequences
Protein-like sequences in cubic lattice
Degeneracy
29
Summary

• Bioinformatics is a promising branch resulted
  from the marriage of computer science and biology
• The exponatioal growth of biological data
  requires machine analysis
• Prediction problems are still open especially in
  RNA Proteins
• Designing fast algorithms, high ability of
  programming skills and Databases design are the
  main skills required to develope software in
  bioinformatics
• Both statistical and constraint approaches are
  used to tackle problems in bioinformatics

30
References

• Computational Molecular Biology An Introduction
  Computational Molecular Biology An
  Introduction, Peter Clote and Rolf Backofen
• CPSP-web-tools a server for 3D lattice protein
  studies (Martin Mann, Mohamad Rabbath ,Cameron
  Smith, Marlien Edwards, Sebastian Will, Rolf
  Backofen )