Systems Biology Final exam Flashcards
1
Q
Systems biology involves
A
- collection of experimental data
- mathematical models
- computer simulations
- validation with additional experimental data
(all of the above)
2
Q
Which of the simulations have emergent properties?
A
Conway’s game of life
3
Q
Which of the following is true regarding systems biology?
A
Can study systems that have feedforward interactions
4
Q
What is cybernetics?
A
The scientific study of control and communication in animals and machines
5
Q
Why the properties of systems cannot be fully understood merely by drawing diagrams of their interactions?
A
You do not get the qualitative interactions or the details of the interactions
6
Q
Which of the following is true regarding the deterministic models?
A
- the same input and parameters always produce the same output
- they can have complex dynamics
- they can be hard to predict without simulations
7
Q
Sensitivity analysis
A
studies how change in parameters change behavior
8
Q
What advancements have permitted the recent success of systems biology?
A
- increased computational power
- advances on analysis of complex systems
- development of quantitative biology techniques
9
Q
What are S, I, and R?
A
variables
10
Q
Which of the following properties is true in a scale free network?
A
Each node can be reached from any other node through a short path
11
Q
Select the coherent feed forward loops
A
coherent feed loops have the same output (+/+ or + for example) no matter which path you take
12
Q
What are network motifs?
A
Patterns that are seen in biological systems more often then in random networks
13
Q
Double negative feedback loops have what basal level for a and b
A
Both high
14
Q
What is a bistable system?
A
a system with two stable equilibrium states
15
Q
In the context of biological systems what is robustness?
A
The ability to maintain biological functions despite perturbations
16
Q
What are modular networks?
A
Networks that contain densely connected functional subunits sparsely connected between them
17
Q
Modularity can emerge as a result of
A
selective forces to maximize the performance and minimize the cost of interconnections in a network
18
Q
What are the kinetic assumptions needed for applying the law of mass action?
A
- fixed volume
- reaction volume is well stirred
- individual reaction events cause infinitesimal changes in concentration
19
Q
If you need to model spatial pattern in a developing embryo what would be the most appropriate mathematical formalism to use?
A
PDE
20
Q
In a dynamical system what is bifurcation?
A
A qualitative change in the behavior of the system due to a change in the parameters
21
Q
What type of systems are used for Turing’s reaction diffusion equation?
A
spatial patterns
22
Q
What do we need to discretize for solving PDEs numerically?
A
time and space
23
Q
Which of the following statements are true regarding parameter estimation with non linear regression?
A
- multiple solutions are possible
- heuristic methods are available for finding parameters close to the optimal
24
Q
Evolutionary algorithms can be used for parameter estimation and reverse engineering of models. What type of optimization algorithm is an evolutionary algorithm?
A
probabilistic search with multiple elements
25
If we need to model a stochastic system with discrete number of molecule which of the following modeling approaches can we use?
Gillespie algorithm
26
Is noise beneficial or detrimental
both
27
Assuming the gene regulatory signals are not modeling but considered inputs of the systems where does the noise come from in the following system
extrinsic noise
28
The following is a Lotka-Volterra predator prey system. What variables represent the number of predators in the population
y
29
What does the population in an evolutionary algorithm contain?
a set of candidate models
30
What does the fitness quantify in an evolutionary algorithm?
how close the simulation is to the experimental data
31
What type of equations are used in the following model?
ODEs
32
What is an emergent property?
a characteristic of a system that arises from complex interactions among its components rather than being directly attributable to any
33
What other reason beyond high throughput data has allowed the application of systems biology?
increased computational power
34
Why does clustering analysis not reveal causality of regulatory relationships?
because correlation does not imply causation
35
What are some examples of abstraction levels in systems biology?
- molecular scale
- ecosystem scale
- tissue scale
- organism scale
- cellular scale
36
When does modeling become necessary?
when intuition reaches its limits
37
Do simple systems have intuitive emergent properties?
no, they have non intuitive emergent properties
38
What do transcription factors do?
proteins that regulate gene transcription
39
What did Alan Turing suggest about patterns in nature?
certain patterns could be explained through interaction of reaction diffusion
40
What 4 types of things do you need to create a model?
- scope
- data
- qualitative information
- feasibility of model
41
What is a correlative model?
given x values, model predicts y values, but doesn't explain why
42
What is an explanatory model?
shows correlation between groups
43
What is a static model?
no changes over time
44
What is a dynamic model?
changes over time
45
What is a deterministic model?
- no random variables
- same input
- parameters produce same output
- hard to predict
- complex dynamics
46
What is a stochastic model?
- random variables
- same input
- parameters produce different outputs
- results make sense
47
What are 4 things used to design a model?
- variables
- parameters
- constants
- interactions
48
What do model diagnostics help with?
showing what is wrong with the model
49
What does model analysis do?
explains model behaviors
50
What does stability on a graph represent?
reaching a steady state
51
What does sensitivity on a graph represent?
how parameter changes can change behavior
52
Explain the paradox that the more facts we learn the less we understand the process
the more info you have on a topic, the more difficult it is to understand its complexity
53
What are characteristics of the languages used by engineers in comparison with the languages usually used by biologists?
engineering languages are standard and quantitative
54
What are randomized networks?
the nodes in networks with random links are connected with equal probability
55
What are biological networks?
nodes are not connected with equal probability
56
What are properties of biological networks?
- connection sparsity
- distribution of connections
- hubs
57
What is a network motif
a recurring structural feature of a network or system that is found more often than one would expect in a corresponding randomly composed system
58
Where are network motifs found
- internet networks
- biochemical networks
- ecological networks
- electronic circuits
- neural networks
59
What could be the bi-parallel motif indicate in food webs
two species that are predators to the same prey, have the same predator
60
What is bistability
when there are 2 stable steady states in a system
61
Steady state
system does not change
62
Stable steady state
system returns after small perturbation
63
Insatiable steady state
system diverges after a small perturbation
64
Is a feedback loop necessary or sufficient for bi stability or memory
- bi stability - necessary but not sufficient
- memory - necessary and sufficient
65
Is positive autoregulation bistable
it can be
66
Linear autoregulation
one instable steady state
67
Sensitive autoregulation
one stable steady state, one instable steady state
68
Ultrasensitive autoregulation
two stable steady states, one instable steady state
69
What can biological switches do
provide memory
70
In the absence of regulation, how do RNA and proteins react vs genes
RNA/proteins - decay
genes - high basal levels
71
Randomized network
nodes in network with random links are connected with equal probability
72
Biological netwokr
nodes in biological interactions that are not connected with equal probability
73
Connection sparsity
only a very small portion of all possible edges are formed
74
Distribution of biological networks
power law distribution of connections or scale free network
75
Hubs
nodes with a high number of connections
76
Scale free network have
small world behavior
77
Negative auto regulation
stronger promoter -> fast response, controlled steady state
78
Bistability and memory
- switches can provide memory
- response is history dependent
79
Design principles that are prevalent in dynamic biological systems
- network motifs
- modularity
- robustness
- redundancy
80
Components in chemical reaction
- molecular species
- ions
81
Interactions
- chemical binding and unbinding
- reaction catalysis
- regulation of activity
82
Rates of interactions depend of
- concentration of reactants
- physio chemical conditions
83
Kinetic modeling assumptions
- spatial homogeneity
- continuum hypothesis
- Law of mass action
84
Equations in kinetic modeling
ODEs
85
intrinsic noise
- from probabilistic character
- particularly important with low numbers of reacting molecules
- inherent to the dynamics of any genetic or biochemical system
86
Extrinsic noise
from random fluctuations in environmental parameters
87
Modeling of stochastic systems
- Langevin approach
- Gillespie algorithm
88
Parameter estimation for linear systems
the exact solution can be calculated analytically by linear regression
89
Parameter estimation for non linear systems
linear regression can also be used for some nonlinear functions if they permit a mathematical transformation to make them
90
Parameter in linear regression
- one function gits best the data
- analytical methods for finding the optimal parameters
- solution is unique
91
Parameter estimation in non linear regression
- infinite function data perfectly
- no simple methods for finding the optimal parameters
- multiple solutions possible
92
Steady state
a state in which the system does not change
93
Stable steady state
a steady state in which the system returns after a small perturbation
94
Instable steady state
a steady state in which the system diverges after a small perturbation
95
Are there design principles that are prevalent in dynamic biological systems?
- network motifs
- modularity
- robustness
- redundancy
96
Robustness
the ability to maintain biological function despite perturbations
97
modeling chemical reaction networks
- components
- interactions
- rates of interactions
98
Spatial homogeneity
- reaction is well stirred
- reactants equally distributed
- rates independent of position in space
99
Continuum hypothesis
- many molecules of each species present
- concentration varies continuously
- individual reaction events cause infinitesimal changes in concentration
100
Law of mass action
reaction is proportional to probability of a collision of the reactants
101
Enzymatic reactions can be described in general as
binding and catalysis
102
Spatial chemical reaction networks
dynamic behavior that is spatially distributed and modeled with PDEs
103
Intrinsic noise
from the probabilistic character of the biochemical reactions
104
Extrinsic noise
from random fluctuations in environmental parameters
105
The Langevin approach
- stochastic differential equation
- variables are continuous
106
The Gillespie algorithm
- direct simulation of the master equation
- variables are discrete
107
What term did Norbert Wiener define
cybernetics
108
What does cybernetics mean
the scientific study of control and communication in the animal and the machine
109
Why the properties of a system cannot be fully understood by drawing diagrams of their interconnections
- because the behavior of the system depends on the specifics of the interconnections
- because changing the strength of an interconnection can change the behavior of the system
110
Why clustering analysis does not reveal the causality of regulatory relationships
because correlation does not imply causation
111
What are network motifs?
overrepresented patterns of network interconnections as compared to a corresponding random network
112
What could the bi parallel motif indicate in food webs?
two species that are predators to the same prey have the same predator
113
Is the cell too complex to use engineering approaches?
no because the properties of electric things have levels of complexity just as a cell does
114
What is a bistable system?
a dynamical system with two stable states
115
Is feedback necessary or sufficient ingredient for biostability?
necessary
116
What is hysteresis?
the dependence of the system state on its past history
117
What is evolvability?
the capacity of a population to adapt in evolutionary time to novel environments
118
What are modular networks?
networks that contain highly connected clusters of nodes that are sparsely connected to nodes in other clusters
119
What are modularly varying goals environments?
rapidly changing environments that have common subproblems but different overall problems
120
What is one of the main advantages of using modeling based on differential equations compared to pother approaches based on statistics?
models based on differential equations can make predictions beyond the observations gathered
121
In a dynamical system model of a set of mixed chemical species what are the components of the system?
the chemical species
122
In a dynamical system model of a set of mixed chemical species what are the rules of the system?
the chemical reactions and rates at which they occur
123
In a dynamical system model of a set of mixed chemical species what are the states of the system?
the concentrations of the chemical species over time
124
What is the difference between models based on difference equations and models based on differential equations?
models based on difference equations treat time as evolving in discrete steps
models based on differential equations treat time steps becoming infinitely small
125
What premise must be true in a biological system for modeling it with ordinary differential equation?
the system components must be well mixed so their concentrations are spatially homogenous
each of the system components must have sufficiently many elements to be treated as continuous quantities
126
How can we model systems with spatial structures using ODEs?
dividing the system into well mixed compartments
127
What is the appropriate mathematical formalism to model systems where continuous aspects of geometry are important of the well mixed assumption does not hold?
partial differential equations
128
What is the process of model calibration?
the process of finding the parameters so that the model simulations recapitulate the observed data
129
Where can the stochasticity in gene expression arise from?
- splicing of mRNA
- binding of DNA polymerase
- binding of transcription factors
- post translational changes of protein
- protein degradation
130
Why transcription initiation is inherently stochastic?
because binding events are the result of random encounters between molecules some present in small numbers
131
What is the difference between deterministic and stochastic simulation?
stochastic simulation includes random processes such as noise while deterministic simulation does not
132
What conditions are necessary for deterministic and stochastic simulations to produce similar results?
large number of molecules and fast promoter kinetics
133
What is the difference between intrinsic and extrinsic noise?
intrinsic noise comes from factors within the system and extrinsic noise comes from factors outside of the system
134
Why differential equations are ideal for modeling dynamic biological mechanisms?
they can predict precise behaviors under perturbations
135
What do X and Y represent in the original Turing system?
the interactions of the system
136
What method was used for numerically solving the Turing system?
Euler
137
What does the population in an evolutionary algorithm contain?
a set of fitness values
138
What does the fitness quantify in an evolutionary algorithm?
how close a model simulation is to the experimental data
138
What are GRN sub circuits?
they are a part of the network that performs individual regulatory tasks
139
What operation is not part of the main loop in an evolutionary algorithm?
creating an initial population of random models
140
Why is the view that the location of the target gene expression is determined solely by the quantitative value of a morphogen gradient overly simplistic?
the pattern and signal strength are network properties as opposed to a property of individual cis regulatory models
141
Generally how is the topology of embryonic genetic regulatory networks?
they are hierarchal and they are deep
142
What type of genes are found in differentiation drivers?
genes encoding for transcription factors
143
What type of genes are found in differentiation gene batteries?
genes encoding for transcription factors
144
What types of mathematical models are frequently used for GRNs?
differential equations and boolean networks
145
How dynamic are protein levels during development?
most protein levels do not change much
146
Which proteins are most dynamic during development?
proteins that are in low abundance
147
What is an example of low dynamic proteins with high abundance during development?
metabolic enzymes
148
What is an example of high dynamic proteins during development?
tissue specific proteins
149
In addition to concentration levels what is another highly dynamic property of some proteins during development?
post translational modifications
150
Did the dynamics of protein and RNA levels correlate?
some protein levels correlated with their mRNA levels but many did not
151
What are some examples of qualitative modeling approaches?
- logical modeling
- graph models
- petri nets
152
What type of information needed for quantitative models are not needed for qualitative models?
- parameters
- mechanistic descriptions of biochemical processes
- quantitative data
153
Regarding the number and type of edges when is a feedback loop positive and when is it negative?
a feedback loop is positive when it contains zero or an even number of negative edges
it is negative when it contains an odd number of negative edges
154
What is the difference between a graph and a hypergraph?
the edges in a graph connects only two nodes
the edges in a hypergraph can connect more than two nodes
155
What are minimal intervention sets in logical models?
they area the minimum sets of deactivators or constitutive activations that achieve a desires behavior in the system
156
What are the two critical factors hindering the construction of comprehensive computational models??
- there are no single computational method that can explain all complex phenotypes from molecules and their interactions
- we do not know enough about all the molecules and their interactions in any one organism
157
The full global model is
stochastic and dynamic
158
What term and process regarding developmental biology did Turing postulate in 1952?
morphogen and reaction diffusion
159
What type of mathematical model did Turing use?
partial differential equations
160
What can the Turing system explain?
how periodic patterns form
161
Why are Turing systems non intuitive?
because it is based on diffusion which by itself destroys spatial patterns
162
What can Wolpert's concept of positional information explain?
how cells can form spatial patterns using gradient signal
163
What is an example of developmental process that follows Wolpert's concept of positional information?
the gap genes pattern in drosophila
164
Why Wolpert's concept of positional information cannot be called a self organizing system?
because it depends on an external signal producing the morphogen gradient
165
Why Wolpert's concept of positional information became initially more accepted than Turing's reaction diffusion system
- because developing one pattern into another was seen as more plausible than patterns from nothing
- because positional information is conceptually simper and easier to explain than reaction diffusion
166
What was the main reason for the revival of Turings reaction diffusion system?
because scientists started to find specific genes that implemented turing systems
167
What is one reason for which understanding metabolism remains a challenge?
metabolic fluxes cannot be measured directly
