Systems Biology

Front 1

Describe the process of "hypothesis-driven" science.

Back 1

Hypothesis-driven science puts emphasis on producing a hypothesis from prior data or theories first and then performing experiments to refute the hypothesis.

Front 2

Describe the process of "data-driven" or "inductive" science.

Back 2

Data-driven science tries to first collect large amounts of data from high-throughput "screening" experiments in order to analyze and build models or hypotheses from that data.

Front 3

Is the "hypothesis-driven" or "data-driven" approach more descriptive of systems biology? Why?

Back 3

The "data-driven" approach lies closer to the workflow of systems biology. The complexity of biological systems and amount of components make it very difficult to propose system-wide hypotheses. This complexity also means that many thousands of experiments are needed to examine the system fully. Models of the system can then be built from the data.

Front 4

Explain the terms "top-down" and "bottom-up" systems biology.

Back 4

Top-down systems biology aims to create mathematical models of a system directly from experimental data via bioinformatics (data-driven). Bottom-up systems biology builds smaller-scale mathematical relations first, before connecting and integrating them into a model of a whole system (hypothesis-driven).

Front 5

What is a biorefinery? What would be its purpose?

Back 5

Biorefineries are proposed as a way to produce organic chemicals, fuels and materials from different types of biomass (straw, wood, corn, algae and so on). A biorefinery would have the dual purpose of replacing existing oil-based products with renewable alternatives as well as producing new products with interesting or practical properties.

Front 6

What is the greatest challenge of building and implementing a biorefinery? What could systems biology help with?

Back 6

Designing the microbial strains ("cell factories") used to convert the treated biomass into usable products. The cost of the metabolic engineering required to craft a useable organism is very great. Systems biology could potentially help speed up the development process while keeping down the cost using novel technologies.

Front 7

The so called "blockbuster method" used in the pharmacological industry is failing, why?

Back 7

The blockbuster method is dependent on finding targets that work for a vast majority of patients (large volumes sold) while minimizing development costs. However, most of the "easy" targets have already been investigated, and it is getting progressively harder to find new targets or drugs without drawbacks.

Front 8

What is "personalized medicine"? What are its advantages?

Back 8

Personalized medicine takes into account that differences between patients may make them respond to the same drug in different ways. If it is possible to correlate a response to e.g. a genotype, then many drugs that are now rejected because they are harmful to a subset of the population could be approved for use on the patients the drug is not harmful to.

Front 9

What role could systems biology play in developing personalized nutrition?

Back 9

The microbes living in the human gut strongly influence our health and well-being. Systems biology could be used to model the gut microbiome and its responses to different diets. This could allow personalized diets that minimize health risks or aid in treating diseases.

Front 10

What other advances would developing a model of the gut microbiome yield?

Back 10

A model could give clues about suitable biomarkers for a wide range of conditions that can then be tested for in clinical settings. Being able to model how the gut microbiome responds to pharmaceuticals (orally administered or otherwise) would help develop safe and effective new drugs. Perhaps some drugs could even target some microorganism/-s directly.

Front 11

Describe the differences between the "components biology" of the 20th century and the systems biology of the 21st century.

Back 11

The field of biology have been drifting from a focus on components (single proteins, genes, or cascades) to progressively larger interconnected systems. With systems biology, the focus has shifted fully to modeling an organism (or set of organisms) in its entirety.

Front 12

Describe the differences in methods used between systems biology and components biology.

Back 12

Component biology was based on techniques based on analytical chemistry such as transcriptomics, genomics and proteomics. The developments that allow systems biology are increases in throughput for traditional methods (using e.g. robotics) combined with new bioinformatics tools (integrative analysis, in silico modeling).

Front 13

Describe the concept of modularity and hierarchy in biology.

Back 13

There is a clear hierarchy in biology, from genetic information, to components (proteins and RNA), to small-scale modules (e.g. signal cascades or series of reactions), to large-scale modules (e.g. cytoskeleton or sprawling reaction networks) that finally give rise to some phenotype. The modularity is the tendency of components to associate into (somewhat) self-contained modules that have less connections between them than they have connections internally.

Front 14

What are the four principle steps of systemic biology?

Back 14

1. Generate -omics data through experiments.

2. Reconstruct biochemical reaction networks.

3. Supplement the model with additional data (e.g. constraints, filling gaps) and characterize (e.g. topology, sensitivity, noise).

4. Draw conclusions and build hypotheses (e.g. understanding biological process, discovering potential targets).

Front 15

What are "GEMs"?

Back 15

GEM is an abbreviation of GEnome-scale Metabolic models. GEMs are like maps of all (or an important subset of) the biochemical reactions in a cell.

Front 16

What are the fundamental features of GEMs?

Back 16

1. Cellular reactions are chemistry, and GEMs therefore contain chemical and physiochemical information.

2. Reconstructions represent genomic and biochemical knowledge.

3. The model contains a number of constraints (e.g. biological, topological, physiochemical or regulatory).

4. Cells exist in a context, and the GEM allows modeling of response.

5. Mass is conserved, simplifying computation (mass-balance).

6. Cells are exposed to selection pressure, meaning that the model can be easily optimized using an objective function.

Front 17

What characterizes a good and practical model?

Back 17

A good model needs to fit experimental data well, reveal some insight into the phenomena it represents and (if possible) give quantitative information. A model does not need to be perfect (e.g. simulating all interactions in a cell), it just needs to be good enough to be practical. A simpler model is often easier to use.

Front 18

What characterizes direct and indirect interactions between pathways or modules in a model?

Back 18

Direct interactions between pathways exist when the proteins directly influence each other (e.g. through binding or phosphorylation). Indirect interactions exist when proteins in different pathways act on the same resources (e.g. share or compete for the same metabolites).

Front 19

How does chemical processes compare to biological processes when it comes to reaction speed?

Back 19

Chemical processes are almost always much quicker. Since biological processes consist of an aggregate of chemical reactions the speed of biological processes must necessarily be much slower than chemical processes.

Front 20

What impact does scale have on the construction of a model?

Back 20

Biological systems range from molecular processes in a cell (e.g. gene expression, protein folding) to populations of multicellular organisms. The model needs to be very different depending on what it is supposed to represent. In order to be as simple as possible, larger models will often gloss over many details contained in models of smaller systems.

Front 21

Explain the idea of "unknown unknowns" and how it pertains to developing a model.

Back 21

Unknown unknowns are concepts or things we don't know anything about, we do not even know they exist. The most useful of models try to realize the limits of our knowledge and allow for expansion as new experiments make more information available to the scientific community. Practically this is achieved by minimizing the amount of assumptions made.

Front 22

Explain the difference between stochastic and deterministic models.

Back 22

A stochastic model contains some kind of uncertainty or element of randomness, meaning that the same inputs can give slightly different outputs. A deterministic model ignores the uncertainty for an idealized process, so that the same inputs always yield the same output.

Front 23

Are stochastic or deterministic models more appropriate for modeling biological systems?

Back 23

Cellular systems usually appear stochastic (every measurement deviates at least slightly from the ideal case), and a stochastic model could be more accurate. However for populations of cells the average or collective response is usually more interesting, and that average often behaves deterministically. Therefore, the simpler deterministic models are often preferred.

Front 24

That's it for the first lecture. Get a bit of a break (you earned it), and come back when you're ready to continue. Turn over the card for an interesting fact about biology!

Back 24

The platypus (Ornithorhynchus anatinus) has tens of thousands of nerves in their bills that are sensitive to magnetic fields. The platypus uses these to sense the nerve impulses of prey hiding in muddy riverbeds.

Front 25

Describe and give examples of hierarchical and metabolic regulation of a reaction pathway.

Back 25

Hierarchical regulation regulates the amount of protein present by targeting the protein production machinery (e.g. transcription factors, splicing, RNA degradation). Metabolic regulation targets the finished protein (e.g. phosphorylation, inhibition).

Front 26

Explain the process for finding a regulatory element in a DNA sequence.

Back 26

Regulatory elements can, since they are usually so short, lose much of their activity through a single mutation. Therefore, such elements are often very highly conserved. If a short conserved sequence appears before a large number of ORFs, then that sequence is most likely a regulatory element.

Front 27

What is a transcription factor network?

Back 27

A transcription factor network is a consequence of TFs often regulating other TFs. If these interactions are mapped, they can be visualized as a network of interactions.

Front 28

Explain the difference between the transcription factor networks in E. coli and yeast.

Back 28

In E. coli the transcription factor network is much more hirearchical (one TF controls expression of all other TFs), whereas in yeast multiple TFs control the expression of all of the other TFs.

Front 29

What is Chromatin immunopercipitation (ChIP)?

Back 29

ChIP is a method designed to find out where proteins bind to DNA. The method consists of binding proteins to DNA, digesting the DNA not covered by protein, precipitating the protein of interest with antibodies, unbinding, purifying and sequencing the DNA. Sequences that show up more often are more likely to bind the protein of interest.

Front 30

Explain robustness in a biological system.

Back 30

A system is robust if there are multiple pathways performing the same function. If one pathway fails, then another pathway can compensate. However, every pathway has a biological cost. Therefore, biological pathways tend to have additional nonessensial functions that differ between the pathways. This means that the redundancy also yields regulatory fidelity for nonessensial systems.

Front 31

Explain the "point" of having a feed-forward loop in transcriptional regulation.

Back 31

Feed-forward loops allow a a signal to function with a delay. This can be used to ignore sudden transient changes in the environment, and thus allows the organism to conserve resources and energy. Feed-forward loops can also function as a sort of signal amplification mechanism.

Front 32

Describe the processes that determine the concentration of a protein in a cell.

Back 32

Production via ribosomes increases the concentration. Protein degradation as well as dilution due to growth reduce the concentration.

Front 33

Describe the correlation (or lack of correlation) between mRNA concentration and protein concentration.

Back 33

The correlation between mRNA concentration and the concentration of the corresponding protein is, across all proteins, very poor. However, on their own many proteins show an almost perfectly linear relationship between mRNA concentration and protein concentration.

Front 34

Explain the concept of secretory classes.

Back 34

Any given protein can have a finite number of secretory signals. The combination of those signals determine the final fate of the protein (secreted, membrane bound etc). In general, most organisms only use a small subset of the theoretically possible classes.

Front 35

Explain the types of regulation the enzyme HMG-CoA Reductase (HMGR) is affected by.

Back 35

Transcriptional regulation through cholesterol (activation) and mevalonate derivatives (repression). Activation through phosphorylation by AMPK (stimulated by mevalonate) and inhibited through dephosphorylation by phosphatase (triggered by insulin). Proteolytic regulation triggered by sterols.

Front 36

Break time! Then continue with lecture three: "Introduction to metabolism".

Front 37

Describe the "bow-tie" structure of metabolism.

Back 37

Biological systems tend to have a large variety in the kinds of raw materials they ingest. These raw materials are then broken down into a smaller number of intracellular metabolites, which are then used to synthesize a large range of molecules. A bow tie is smaller in the middle. Many substrates -> few intermediaries -> many products.

Front 38

Describe the relation between connectivity and regulation in a metabolic network.

Back 38

A high degree of connectivity requires that the network can be regulated on a global level in order to be efficient. A high degree of connectivity even imposes global regulation, since all parts are dependent on many others to function. Changing one part will necessarily effect a change on the other parts.

Front 39

Describe the different kinds of transmembrane transport processes that take place in a cell.

Back 39

Passive: simple diffusion (through the membrane) or facilitated diffusion (via channel proteins. Can only work down the concentration gradient.

Active: requires energy (e.g ATP), symport or antiport (transport of something else down a concentration gradient).

Front 40

Dsecribe, in general terms, the central carbon metabolism.

Back 40

Highly conserved between species. Usually consists of glycolysis (breakdown of sugars), TCA cycle (or some variation thereof, cyclic breakdown of energy source), some fermentative pathways/pathways and oxidative phosphorylation (e.g. electron transport chain). Usually provides redox power, precursor metabolites and usable Gibbs free energy.

Front 41

What is the Crabtree Effect?

Back 41

The Crabtree Effect is the metabolic shift that yeast undergoes when it is grown on an abundance of glucose. The abundance of glucose triggers repression of repression leading to fermentation.

Front 42

Describe the different families of amino acids in E. coli.

Back 42

Glutamate family, starts with ketoglutarate from TCA cycle. Pyruvate family, starts with pyruvate. Serine family, starts with 3-phosphoglycerate from glycolysis. Aspartate family, starts with oxaloacetate from TCA cycle. Aromatic family, starts with phosphoenolpyruvate and erythrose-4-P. The aromatic and aspartate families are interesting for industrial applications. Histidine is not synthesised as part of any family.

Front 43

Describe why lipid metabolism is so highly regulated.

Back 43

Lipids are a very dense but comparatively slow kind of energy storage. It is inefficient to use it to store energy short-term. Because of this, the cell needs to be sure of abundance for the foreseeable future before it starts to synthesize lipids.

Front 44

Break. Fourth lecture ("Network Reconstruction") coming up!

Front 45

What role does substrate specificity play when reconstructing a metabolic network?

Back 45

In order to describe the network of reactions, one needs to be able to specify what those reactions are. Therefore, the specificity of the enzymes are very important to find out so that all reactions that can take place are present, while those that cannot take place are not. This information can be very difficult to find.

Front 46

How is charge balance taken into account when designing a metabolic network?

Back 46

In general, simpler networks ignore charge balance. In the more advanced networks charge is incorporated by describing the physiological conditions for the reactions, calculating the pKa value for each functional group and determining the charge of a molecule in such conditions. Information about charge is then used to elucidate stoichiometry and thermodynamics for the reactions.

Front 47

How is directionality considered for metabolic networks?

Back 47

If a reaction is irreversible, then that can have an important effect on the network. However, finding the information can be difficult. The Gibbs free energy of reaction can be very different at physiological condition from what is expected. If the directionality cannot be found, the reaction is usually assumed to be reversible.

Front 48

Explain localization in the context of metabolic networks.

Back 48

Localization of proteins predicted by software (e.g. PSORT and SubLoc) or experimentally predicted (HTS using GFP-tagging of proteins. If no data is available, proteins are assumed to reside in the cytosol. if faulty, could produce gaps in network, misrepresent the properties of the organism or give rise to erroneous conclusions.

Front 49

Explain why the relation between genes and rections aren't 1-to-1.

Back 49

Isozymes: Some reactions are catalyzed by more than one enzyme.

Multifunctional enzymes: Some enzymes catalyze more than one reaction.

Complexes: Some reactions are catalyzed by complexes, requiring more than one gene.

Front 50

What is the S-matrix?

Back 50

The S-matrix is a record of all of the stochiometric coefficients for all metabolites and reactions. Metabolites are recorded as rows and reactions as columns. The S-matrix is usually very sparse (many elements = 0).

Front 51

Describe a general workflow for the construction an use of a metabolic model.

Back 51

Use annotated gene sequence data from databases to draft a reconstuction. Curate the reconstruction using known reactions, biochemical and genetic data. Describe common inputs to and outputs from the model with detailed physiological data. Validate and develop the model by testing and comparing with -omics data from experiments. Use the model for knowledge generation.

Front 52

You know what time it is! Break! Fifth lecture next: "Metabolism of E.coli and other prokaryotes".