Higher Order Systems

1

• Higher Order Systems

2
In this presentation

• Part 1 Genetic Regulatory Networks
• Part 2 Molecular Pathways
• Part 3 Protein Interactions
• Part 4 Modeling Regulatory Networks

3
Part1

• Genetic Regulatory Networks

4
Genetic regulatory networks
5
Higher order systems

• Although genes and proteins can be studied
  individually, more insight into their functions
  can be gained by studying higher-order systems,
  that is, molecular pathways and networks, cells,
  tissues, organs and whole organisms
• This allows their physical and functional
  interactions to be determined in the widest
  possible context

6

• The work of Tavazoie et al (1999) is vividly
  known for systematic determination of genetic
  network architecture
• Cell signaling pathways are linked to genetic
  regulatory pathways in ways we are just beginning
  to unscramble
• The most enormous bioinformatics project in front
  of the scientists is unscrambling this regulatory
  network, which controls cell development from the
  fertilized egg to the adult

7

• It would become possible to know which gene to
  perturb or which sequence of genes to perturb,
  and in what order to guide a cancer cell to
  nonmalignant behaviour or to apoptosis
  programmed to cell death
• Or to guide the regeneration of some tissues, so
  that if someone has lost half of the pancreas,
  the damaged portion could be regenerated
• Or to regenerate the beta cells in people who
  have diabetes

8

• Suppose about 10 genes are picked out that are
  known to regulate one another, then a circuit
  could be built about their behaviour. It is a
  good thing and one should do this but the down
  side will be that those 10 genes have inputs from
  other genes outside that circuit. Therefore, it
  is like taking a little chunk of the circuitry
  that is embedded in a much larger circuit of
  thousands of genes in it. The behaviour can not
  be then properly assessed as to how and what
  impact the outside genes would create

9

• It is known for years that every neuron in the
  lobster gastric ganglia a nerve bundle going to
  the animals digestive system, all the synaptic
  connections and the neurotransmitters
• There would be 13 or 20 neurons in the ganglion
  and still its behaviour cannot be figured out
• No mathematician would ever think that
  understanding a system with 13 variables is going
  to be an easy thing to do

10

• In the human genome case, there would be more
  than 100,000 variables i.e. there would be
  2100,000 states, which is roughly 1030,000
• So even if genes are treated to be on or off,
  there are 1030,000 states (which is false as
  genes show graded level of activity)
• It is mind boggling because the number of
  particles in known universe is 1080

11
Types of pathways

• Molecular pathways
• Metabolic pathways
• Signaling and regulatory pathways
• Protein interaction networks

12
Part2

• Molecular Pathways

13
Representation of pathways and networks

• Molecular pathways and networks can be
  represented by graphs, with molecules at the
  nodes and relationships shown by links
• In metabolic pathways, nodes represent substrates
  or intermediates and links represent their
  catalytic interconversion by enzymes
• In signaling and regulatory pathways, nodes
  represent proteins and links indicate the
  transfer of information
• Graphs of molecular pathways are generally
  directional and can show positive and negative
  interactions

14
Reconstruction of molecular pathways

• Pathways and networks can be mapped directly by
  substrate feeding experiments and in vitro enzyme
  assays
• More recently, a number of indirect but
  high-throughput methods have been developed
  thanks to the advent of functional genomics
• These methods include pathway reconstruction from
  expression data, protein interaction and
  comprehensive mutagenesis programs

15
Modeling molecular pathways

• Mathematical models of biochemical reactions are
  often based on differential equations that
  predict the change in concentration of particular
  molecules over time
• Simultaneous differential equations can be used
  to model entire pathways and several software
  applications are available for this task,
  including GEPASI and BioQuest

16

• There are limitations to the use of simultaneous
  differential equations and these have been
  addressed through the development of stochastic
  models based on the Gillespie algorithm, which is
  incorporated into programs such as StochSim

17
Subgraph with main interactions between GAD and
GABA-receptors, derived from the linear model. P.
D'haeseleer, X. Wen, S. Fuhrman, and R. Somogyi
(1999) Linear Modeling of mRNA Expression Levels
During CNS Development and Injury
18
Overview of Procedures for Preparing and
Analyzing Microarrays of Complementary DNA (cDNA)
and Breast-Tumor Tissue. As shown in Panel A,
reference RNA and tumor RNA are labeled by
reverse transcription with different fluorescent
dyes (green for the reference cells and red for
the tumor cells) and hybridized to a cDNA
microarray containing robotically printed cDNA
clones. As shown in Panel B, the slides are
scanned with a confocal laser scanning
microscope, and color images are generated for
each hybridization with RNA from the tumor and
reference cells. Genes up-regulated in the tumors
appear red, whereas those with decreased
expression appear green. Genes with similar
levels of expression in the two samples appear
yellow. Genes of interest are selected on the
basis of the differences in the level of
expression by known tumor classes (e.g.,
BRCA1-mutationpositive and BRCA2-mutationpositiv
e). Statistical analysis determines whether these
differences in the gene-expression profiles are
greater than would be expected by chance. As
shown in Panel C, the differences in the patterns
of gene expression between tumor classes can be
portrayed in the form of a color-coded plot, and
the relations between tumors can be portrayed in
the form of a multidimensional-scaling plot.
Tumors with similar gene-expression profiles
cluster close to one another in the
multidimensional-scaling plot. As shown in Panel
D, particular genes of interest can be further
studied through the use of a large number of
arrayed, paraffin embedded tumor specimens,
referred to as tissue microarrays. As shown in
Panel E, immunohistochemical analyses of hundreds
or thousands of these arrayed biopsy specimens
can be performed in order to extend the
microarray findings.
19

• The two basic clusters of a) early and b) late
  upregulated genes as identified by percolation
  clustering. Color coding of the expression
  profiles is as follows black means gene
  expression is the same as it was at 2 hours of
  development increasing tint of red color means
  increasing expression relative to 2 hours and
  increasing tint of green color means decreasing
  expression relative to 2 hours
• The bottom portions of the figure display
  expression profiles of the corresponding genes
  the red curves are the mean expression. Only
  genes whose connectivity to the cluster origins
  is greater than 20 were included in these plots.

20
 
Templates for Looking At Gene Expression
Clustering By Daniel B. Carr, Roland Somogyi and
George Michaels
21
Gene co-expression pairs in CNS development and
injury
22
Mutual information tree for genes expressed in
rat spinal cord. Michaels G, Carr DB, Wen X,
Fuhrman S, Askenazi M, Somogyi R (1998) Cluster
Analysis and Data Visualization of Large-Scale
Gene Expression Data
23
Gene expression waves. (a) Normalized gene
expression trajectories from Fig. 2 are shown
grouped by waves determined by Euclidean
distance clustering. Graphs show average
normalized expression pattern or wave over
the nine time points for all the genes in each
cluster (the time of birth is marked by a
vertical line). Within each wave, genes are
grouped according to gene families, not according
to proximity as determined by Euclidean distance.
(b) Euclidean distance tree of all gene
expression patterns (for annotated tree, see
http//rsb.info.nih.govymol-physiolyPNASytree.html
). Major branches correspond to waves in a. (c)
Plots of all normalized time series, highlighting
wave 3 (Left, white lines) and a subcluster of
wave 3 (Right, white lines plotted on top of
remaining genes of wave 3 in red). Subclusters
(secondary branching) were selected by visual
inspection from tree in b e.g., the plotted time
series of the wave 3 subcluster correspond to
branchlet highlighted in white within wave 3 in
b. (d) PCA. Principal components projection
viewed as a three-dimensional stereo plot. Each
point mapped in three-dimensional space
represents an expression time series
corresponding to a gene in Fig. 2. Highlighted
points correspond to Euclidean distance wave 3
(red triangles), wave 4 (green squares), and the
remaining genes (blue octagons)
24
Molecular pathway resources

• There are many resources for viewing molecular
  pathways on the Internet
• One of the most comprehensive for metabolic
  pathways is KEGG and this also shows a selected
  range of regulatory pathways
• An important feature of such resources is that
  the contents of the maps are integrated with
  other databases by way of hyperlinks

25
Part3

• Protein Interactions

26
Interactions and pathways

• Proteins that physically interact with each other
  may be involved in the same molecular pathway or
  network, or may form part of a multi-subunit
  complex
• Using this principle, pathways can be
  reconstructed based on evidence of protein
  interactions
• However, information from other sources e.g.
  gene expression patterns and mutant phenotypes
  may also be useful

27
Handling Y2H data

• Yeast two-hybrid (Y2H) screens produce large
  amounts of protein interaction data, but there is
  a relatively high level of spurious results
  (false positives and false negatives)
• This problem can be addressed by scoring
  interactions for reliability, based either on the
  repeatability of interactions over multiple
  experiments, or by the number of times a given
  bait will trap independent clones representing
  the same prey
• Even so, similar large-scale screens tend to
  identify different (although) overlapping sets of
  interactions

28
Protein interaction databases

• Several databases have been set up to store the
  interaction data arising from large-scale Y2H
  screens
• However, much more information on protein
  interactions is available in the scientific
  literature and a current challenge in
  bioinformatics is the assimilation of these
  interaction data from diverse sources

29
The interactome

• It is sum of all protein interactions in the cell
• The simplest way to represent protein
  interactions is a graph with proteins as nodes
  and interactions as links
• However, when large numbers of proteins are
  considered, the graphs become too complex
• They can be simplified by clustering functionally
  similar proteins, resulting in a functional
  interaction map that links fundamental cellular
  processes

30
Part4

• Modeling Regulatory Networks

31
The cell

• It can be regarded as a compartmentalized set of
  molecular pathways and networks distributed in
  space and restricted by membranes
• Any model of a cell must incorporate these
  features
• A useful modeling resource is Virtual Cell, in
  which the cell is defined as a collection of
  structures, molecules, reactions and fluxes
• The user can define biological or mathematical
  models for cell function

32
Modeling tissues and organs

• Tissues and organs comprise organized population
  of interdependent cells
• Modeling depends on an accurate description of
  the geometry of the tissue and must include any
  time-dependent processes
• For example, modeling the heart requires a
  description of its anatomy and the way in which
  action potentials are propagated
• The model must take into account the fact that
  cardiac muscle is an anisotropic system

33
Modeling organisms

• In order to model an entire organism, it is
  necessary to have a sound understanding of the
  principles underlying development
• For most multicellular organisms there is too
  little information and the developmental program
  too complex for this to be achieved

34
Nematode C. elegans modeled

• The nematode has a number of features that make
  it an ideal system upon which to base a
  developmental model
• It is a simple organism (it has about 1000
  somatic cells) whose somatic cell lineage is
  invariant, making perturbations in development
  very easy to identify
• The genome has been sequenced indeed, it was the
  first genome of a multicellular organism to be
  sequenced
• It also relatively easy to study the physiology
  of this organism, and hence a complete wiring
  diagram of C. elegans nervous system is available

35
Modeling spaces

• Models of C. elegans development have been
  generated based on the concept of three spaces
• Genomic space
• Cellular space
• Developmental space

36
Relationships among three spaces