Computational Systems Biology

1
Computational Systems Biology

• Prepared by
• Rhia Trogo
• Rafael Cabredo
• Levi Jones Monteverde

2
What are Biological Systems?

• Popular Notion
• It is a complex system consisting of very many
  simple and identical elements interacting to
  produce what appears to be complex behavior
• Example Cells, Proteins

3
What are Biological Systems?

• Realistic Notion
• It is a system composed of many different kinds
  of multifunctional elements interacting
  selectively and nonlinearly with others to
  produce coherent behavior.

4
What are Biological Systems?

• Complex systems of simple elements have functions
  that emerge from the properties of the networks
  they form.
• Biological systems have functions that rely on a
  combination of the network and the specific
  elements involved.

5
Molecular vs. Systems Biology
Biology

• In molecular biology, gene structure and function
  is studied at the molecular level.
• In systems biology, specific interactions of
  components in the biological system are studied
  cells, tissues, organs, and ecological webs.

6
From Systems Biology to Computational Biology

• Biological Systems are complex, thus, a
• combination of experimental and
• computational approaches are needed.
• Two Branches of Computational Biology
• Knowledge Discovery (Data mining)
• Simulation-based Analysis

7
Knowledge Discovery

• Extracting hidden patterns from a large quantity
  of data forming a hypothesis
• Steps
• Data selection
• Data cleaning
• Transforming to a Data Mining technique
• Data Mining Technique
• Interpretation

8
Problems of Knowledge Discovery

• Too much data!
• Solution
• use heuristics
• use Hidden Markov Model

9
Hidden Markov Model

• Used in finding the protein structure from the
  sequence
• Hidden Markov Model (HMM)-based search methods
  makes use of position-dependent scores to
  characterize and build a model for an entire
  family of sequences

10
Simulation-based Analysis

• Simulation-based analysis tests hypotheses with
  in silico experiments, providing predictions to
  be tested by in vitro and in vivo studies.
• faster and more economical.
• Example Folding_at_Home

11
Folding_at_Home

• Simulates protein folds
• Folds dictate the function of the protein
• Unfolding was discovered by Christian Anfinsen
• When folds do not fold properly, it leads to
  diseases such as Alzheimers disease, Mad Cow,
  Parkinsons disease
• If the fold of the protein is known then it can
  also be unfolded

12
Folding_at_Home

• Runs on a distributed system
• Runs as a screensaver
• Downloadable at
• http//folding.stanford.edu

13
Databases and Tools

• Languages
• Systems Biology Markup Language
• CellML
• Systems Biology Workbench
• Databases
• Kyoto Encyclopedia of Genes and Genomes
• Alliance for Cellular Signaling
• Signal Transduction Knowledge Environment

14
p53

• Protein 53
• Produces 53 proteins kiloDaltons
• Guardian of the genome
• Detects DNA damages
• Halts the cell cycle if damage is detected to
  give DNA time to repair itself

15
p53

• If (damage equals true and repairable true)
• halt cell cycle
• else
• if(damage equals true and repairable false)
• induce apoptosis

16
The Cell Cycle

• G1 - Growth and preparation of the chromosome
  replication
• S - DNA replication
• G2 - Preparation for Mitosis
• M - Chromosomes separate

17
Checkpoints for DNA Double Strand Breakage
18
Cancer Cell Network
19
p53
activates
deactivates
p53
p21
CDK
No cell cycle!
20
Cancer Drugs

• Alkylating agents
• Antimetabolites
• Vinca alkaloids
• Taxanes
• all inhibit the cell cycle

21
Properties of a Drug

• Absorption
• Distribution
• Metabolism
• Extraction
• Toxicology

22
ADME/Tox

• Target selection
• Proteomics and Genomics help
• Prediction
• Comparison of Prediction
• Validation

23
Optimization

• Eliminate leads that could lead to failure
• Kill early
• There is a danger that possible good leads might
  be killed
• Save time
• Kill late
• All possibilities are explored
• expensive

24
p53
25
Robustness in Biological Systems
26
The Cost of Robustness

• Robustness is not a good characteristic for all
  types of cells.
• Example The robust cancer cell!
• Systems that are robust against common
  perturbations are often fragile to new
  perturbations (vulnerability of complex networks)

27
Advantages of Computational Systems Biology

• It is highly relevant in discovering more complex
  relationships involving multiple genes
• This may create new opportunities for drug
  discovery
• Better medical therapies for individual treatments

28
Whats to come?

• Current work is on small sub-networks within
  cells.
• Feedback circuit of bacteria chemotaxis
• Circadian Rhythm
• Parts of signal-transduction pathways
• Simplified models of the cell cycle
• Models of the Red blood cells

29
Whats to come?

• Research has begun on larger-scale simulations
• Biochemical network level
• Simulation of Epidermal Growth Factor (EGF)
  signal-transduction cascade
• The Physiome Project

30
Biochemical Networks

• Problem
• The behavior of cells is governed and
  coordinated by biochemical signaling networks
  that translate external cues (hormones, growth
  factors, stress, etc.) into adequate biological
  responses such as cell proliferation,
  specialization or death, and metabolic control.
• Motivation
• Deep understanding of cell malfunction is
  crucial for drug development and other therapies.

31
Biochemical Networks
32
Biochemical Networks
33
Interpreting Biochemical Networks as Concurrent
Communicating Systems

• Biochemical networks are analogous to concurrent
  computer systems in many respects.
• Concurrent systems are built up using basic
  concepts such as choice, recursion, modularity,
  synchronization, and mobility.
• By exploiting these analogies, the existing tools
  and formalisms for computing systems can be
  applied to biochemical networks.

34
Concurrency Theory

• Concurrent, communicating systems have been the
  subject of intense study by Computing Scientists.
  Rich theories and tools have been developed to
  aid in design, analysis and verification of such
  systems.
• Concurrent systems are inherently complex. To
  manage complexity, theories and tools have been
  developed to allow programmers to simulate
  behaviour. Simulators allow the analysis of
  traces through concurrent executions and provide
  a testbed for experimentation.
• At a more abstract level, temporal analysis
  involves proving that a concurrent system adheres
  to a temporal property, i. e. it can be shown
  that a network protocol always delivers data
  packets in the same order they were sent.

35
Concurrency

• A concurrent system is one where multiple
  processes exist at the same time. These processes
  execute in parallel and potentially interact with
  each other. As an example of a concurrent
  system, consider an internet banking site. The
  server and multiple client processes exist at the
  same time, with interactions occurring between
  the individual clients and the server.

36
Concurrency in Biochemical Networks
Biochemical networks are also concurrent
communicating systems. Pathways consist of
sequences of interactions which sometimes affect
other parallel pathways. As an example, consider
two pathways involved in cell division. The Ras-
Raf pathway which triggers the cell division and
the PI- 3K- Akt pathway which keeps the cell
alive are both triggered by the same growth
factor. The sequences of interactions in both
pathways run concurrently with some interaction
i. e. Akt inhibits Raf.
37
The Physiome Project

• A worldwide effort to define the physiome by
  developing databases and models which will
  facilitate the understanding of the integrative
  functions of cells, organs and organisms.
• def. Physiome is the quantitative and integrated
  description of the functional behavior of the
  physiological state of an individual or species.

38
The Physiome Project

• Main Objective
• to understand and describe the human
  organism, its physiology and pathophysiology
  quantitatively, and to use this understanding to
  improve human health.

39
The Physiome Project

• Specific Objectives
• To develop and database observations of
  physiological phenomenon and interpret these in
  terms of mechanism (a fundamentally reductionist
  goal).
• To integrate experimental information into
  quantitative descriptions of the functioning of
  humans and other organisms (modern integrative
  biology glued together via modeling). 
• To disseminate experimental data and integrative
  models for teaching and research. 

40
The Physiome Project

• Specific Objectives
• To foster collaboration amongst investigators
  worldwide, in an effort to speed up the discovery
  of how biological systems work. 
• To determine the most effective targets
  (molecules or systems) for therapy, either
  pharmaceutic or genomic. 
• To provide information for the design
  tissue-engineered, biocompatible implants.

41
The Physiome Project

• Issues being addressed
• Markup language
• -- development of SBML (in Caltech) for
  representing biochemical networks and CellML for
  electrophysiology, mechanics, energetics and
  general pathway.
• Mathematical models
• -- development of models that are anatomically
  based and biophysically based to link gene,
  protein, cell, tissue ,organ and whole body
  systems physiology.

42
The Physiome Project

• Issues being addressed
• Web-accessible databases
• -- For easy data exchange, groups at MIT and
  UCSD are developing standards for this.
• Example databases Genomic Databases,
  Protein Databases, Material Property Databases,
  Anatomical Model Databases, Clinical Databases
• Development of new instrumentation
• Development of Modeling tools, GUIs and
  web-accessible tools for visualization of complex
  models.

43
The Physiome Project

• Sub-Projects
• Microcirculation
• A common functional system between organs It
  provides an important coupling between cells,
  tissues, and organs.
• Available online http//www.bme.jhu.edu/news/m
  icrophys

44
The Physiome Project

• Sub-Projects
• Musculo-skeletal system
• Continues to extend the database of
  parameterised bone geometry to individual
  muscles, ligaments and tendons.
• Available online http//www.bioeng.auckland.ac
  .nz/projects/nerf/skeletal.php

(a) (b) (a) Anatomically
detailed model of Skeleton. (b) Rendered finite
element mesh for the bones and a subset of the
muscles
45
The Physiome Project

• Sub-Projects

(a)
(b) Computational model of the skull and torso.
(a) The layer of skeletal muscle is highlighted.
(b) The heart and lungs shown within the torso.
46
The Physiome Project

• Sub-Projects
• Cardiome Project
• An attempt to provide an integrated model of the
  heart, incorporating electrical activation,
  mechanical contraction, energy supply and
  utilization, cell signaling and many other
  biochemical processes.

Heart model with a textured epidermal surface
47
The Physiome Project

• A) Heart Structure

(a) (b)
(c) Fibrous-sheet architecture of the heart.
Ribbons are drawn in the plane of the myocardial
sheets (a) on the epicardial surface of the
heart, (b) at midwall, and (c) on the endocardial
surface. Note the large fibre angle changes.
These fibre-sheet material axes are needed for
computation of both myocardial activation and
ventricular mechanics.
48
The Physiome Project

• A) Heart Structure

The finite element model of the right and left
ventricle of the heart showing various anatomical
structures. Geometric information is carried at
the nodes of the finite element mesh and
interpolated with cubic Hermite basis functions.
49
The Physiome Project

• B) Ventricular Mechanics

Mechanics of the cardiac cycle, computed by large
deformation finite element analysis, at (a) zero
pressure state, (b) end-diastole, (c)
mid-systole, (d) end-systole. Note the apex to
base shortening and the twisting about the long
axis. Also note the six generations of discretely
modeled coronary vessels embedded within the
myocardial elements which are used to compute
coronary flow throughout the cardiac cycle.
50
The Physiome Project

• B) Ventricular Mechanics

The collagenous structure of the extracellular
myocardial tissue matrix, as revealed by confocal
microscopy. The material axes used for defining
mechanical and electrical constitutive laws in
the continuum modeling of the myocardium are
based on these microstructurally defined axes.
51
The Physiome Project

• C) Myocardial Activation

Activation wavefront computed on the finite
element model using finite difference techniques
based on grid points which move with the
deforming myocardium. Bidomain current
conservation equations are solved with
transmembrane ionic currents. The stimulus in
this case is a point on the left ventricular
endocardial surface near the apex. The activation
sequence is heavily influenced by the
fibrous-sheet architecture of the myocardium.
52
The Physiome Project

• D) Coronary Perfusion

E) Ventricular Fluid Flow F) Human Torso model
has been developed which includes the heart,
lungs and the layers of skeletal muscle, fat and
skin. Current flow from the heart into the torso
is computed in order to predict the body surface
potentials arising from activation of the
myocardium.
Computed flow in the coronary vasculature
53
The Physiome Project

• Sub-Projects
• Lungs
• Development of models of the integrated function
  of various physical processes operating in the
  lung.

• Bladder and Prostate
• An anatomically detailed model of the bladder
  and prostate is developed.
• Circulation System
• A model of the circulation system is being
  developed based on the Visual Human Project
  dataset (http//www.nlm.nih.gov/research/visible)

54
Whats to come?

• Development of Precision Models
• Simulation requires the integration of multiple
  hierarchies of models that have different scales
  and qualitative properties
• Some biological processes take place within
  milliseconds while others may take hours or days
• Example Protein folding vs. Cell Mitosis

55
Whats to come?

• Development of Precision Models
• Biological processes can involve the interaction
  of different types of processes
• (i.e. biochemical networks coupled to protein
  transport, chromosome dynamics, cell migration or
  morphological changes in tissues)

56
Whats to come?

• Development of Precision Models
• Types of modeling
• Using differential equations and stochastic
  simulation
• Many cell biological phenomena require
  calculation of structural dynamics
• Deformation of elastic bodies
• Spring-mass models and other physical processes

57
Resources

• Kitano, H. . Computational Systems Biology .
  Nature, (420) . pp. 206 210. November 2002.
• p53 Mutation Database Analysis and Search.
  Available Online http//p53.genome.ad.jp .
• Kodratoff, Y. About Knowledge Discovery in
  Texts A Definition and an Example . 2000 .
  Available Online http//www.lri.fr/ia/articles/
  yk/2000/kodratoffupb.pdf .
• The Cell Cycle . Available Online
  http//users.rcn.com/jkimball.ma.ultranet/BiologyP
  ages/c/CellCycle.html
• Head-Gordon, T. and Wooley, J. Computational
  Challenge on Structural and Functional Genomics .
  IBM Journal ,(40,2) . pp. 265-300. 2001.
  Available Online http//www.research.ibm.com/jo
  urnals/sj/402/headgordon.pdf.

58
Resources

• Larson, S. Folding_at_Home and Genome at Home in
  Distributed Systems. Available Online
  http//www.stanford.edu/smlarson/ppt/Carleton_Mar
  ch01.ppt
• Folding_at_home. Available online
  http//folding.stanford.edu .
• Comparative Visualization of Protein
  Structure-Sequence Alignments . Available
  Online http//www.cse.ucsc.edu/research/slu
  /lego.html .
• The Bioengineering Institute, Heart Physiome.
  Available Online http//www.bioeng.auckland.ac.
  nz/projects/heart/heart.php. (Feb 2003).
• The Biology Project-Cell Biology . Available
  Online http//www.biology.arizona.edu/call-bio/
  tutorials/cell-cycle/cells3.html.

59
Resources

• Cancer Biology Online . Available Online
  http//www.iona.edu/faculty/csackernon/cancer/p53/
  p53-2.htm.
• Cancer Drugs General Information-How Do Cancer
  Drugs Work? Available Online
  http//www.bccancer.bc.ca/PPI/CancerTreatment/Canc
  erDrugsGeneralInformationforPatients/DrugsWork.htm
  
• Genome Remodelling in Mammalian Cells.
  Available Online http//pingu.salk.edu/wahl/mi
  ssions.html.
• ADME/Tox. Available Online
  http//www.genego.com/tutorial/index.shtml?23
• Breneman, C..ADME PropertyPrediction. Available
  Online http//www.rpi.edu/locker/82/001182/publ
  ic_html/files/presentations/MACC_Caco2New5/tsld005
  .htm
• ADME/TOX Related Information. Available
  Online http//www.caddininformatics.com/PPT/sld
  006.htm.

60
Resources

• Ideker, T. et al. Integrated genomic and
  proteomic analyses of a systematically perturbed
  metabolicz network. Science 292, 929934 (2001).
• Borisuk, M. T. Tyson, J. J. Bifurcation
  analysis of a model of mitotic control in frog
  eggs. J. Theor.Biol.
• Chen, K. C. et al. Kinetic analysis of a
  molecular model of the budding yeast cell cycle.
  Mol. Biol. Cell 11, 369391 (2000).
• Edwards, J. S., Ibarra, R. U. Palsson, B. O.
  In silico predictions of Escherichia coli
  metabolic capabilities are consistent with
  experimental data. Nature Biotechnol. 19,
  125130 (2001).
• Alon, U. et al. Robustness in bacterial
  chemotaxis. Nature 397, 168171 (1999).
• Barkai, N. Leibler, S. Robustness in simple
  biochemical networks. Nature 387, 913917 (1997).

61
The End! Thank you