Justification
In a research project such as a PhD thesis, it is usual to follow a linear path.
Observation, hypothesis, predictions, experimentation, conclusions.
Then you make new observations, think of new hypothesis and so on and so forth.
We call that scientific method.
Its elegance and power is virtually unbounded.
Reality, however, is usually messier.
Research is not a path, it’s a maze.
You have some ideas, you make some experiments and you obtain some findings.
Most of the time nature pushes you around, blocks your path, laughs at your preconceived ideas and challenges you relentlessly.
This thesis does not fight that apparent chaos, it embraces it.
Famous scientist Uri Alon wrote in Alon 2007 that sometimes it is more interesting to follow the path lead by research itself which is convoluted, not linear.
Hopefully, at the end of it, we will be able to connect the dots of the work done, just like Steve Jobs said in his famous commencement speech in Stanford University in 2005.
Since one can only connect the dots backwards, this will be done at the end of this thesis, in the Conclusions section.
As it has been previously mentioned in the Abstract, this thesis in devoted to the study of biological networks at the molecular level.
The work presented here is the result of the collusion of two large fields of knowledge: complex systems and molecular biology.
The network is the quintessential complex system: large amount of components, very intricate structure and highly dynamic.
Network analysis has been so successful in the last two decades because is one of the disciplines that is better suited to explain the reality that surround us: highly interconnected, everchanging, global, de-localized and apparently unpredictable.
On the other hand, molecular biology is with no doubt the most promising branch of science.
Its possibilities are endless: personalized health-care, bioenergetic production, bioremediation, industrial biotechnology and many more world-changing applications can be drawn from manipulation of living organisms.
However, to achieve those goals we first have to understand these very complex living systems.
This thesis aims to contribute at that exact point.
To develop a series of tools and novel analyses to improve our understanding of those organisms, always revolving around the concept of network.
Thesis outline
As it was briefly mentioned in the Abstract, this thesis is divided in five main parts.
An extensive literature review about systems biology and network theory is carried out in the first chapter, Chapter 2. Chapters 3, 4 and 5 represent the core of the work done.
Since this thesis is devoted to network analysis, it seems appropriate to illustrate the structure of these main chapters with a network representation (Figure 1.1).
This figure is structured representing in each level one of the axes defined in the abstract: the biological one at the top, the network one in the middle and the mathematical one at the bottom.
The work done in each chapter is a combination of different elements in those axes.
Chapter 3 analyses the topology of protein-protein intereaction networks of viruses of the Potyvirus genus with some of their main plant hosts.
The interactions between both organisms are studied in detail.
Besides, the topology of the virus is used as a channel to explain the relationship between genotype and phenotype through organismal fitness.
Chapter 4 is focused on the comparison of different constraint-based methods of metabolic flux determination at steady state in E. coli.
Its main characteristics are evaluated and the validity of their assumptions is address and quantified.
Chapter 5 deals with the mathematical description of metabolic networks and how to make them much more flexible and richer.
The community structure of E. coli is analysed as well.
Contributions
The main contributions of this work are the following:
• (Chapter 3) Novel analysis of protein-protein interaction networks in viralplant dual systems.
Topology, connectivity, similarity, perturbation and effect propagation analyses are proposed to better understand the system.
• (Chapter 3) An approach for using molecular network topology as a connection between genotype alterations (mutants) and phenotype measurements (fitness).
This produces a triple system (genotype-topology-phenotype) that is able to explain the variability of the organism performance.
• (Chapter 4) An exhaustive analysis of the most popular constraint-based methods for obtaining metabolic flux distributions.
Data vs. model agreement, biological plausibility, pathway profiles, solution sizes, growth rate prediction capability, robustness and biological premise quantification are proposed to address the quality and potency of the methods.
• (Chapter 5) Novel graph description of metabolic networks including topology, stoichiometry, directionality and reaction flux information.
• (Chapter 5) Determination of the coarse-grained structure of the E.coli metabolic network under different enviromental conditions.
Robustness, topological analysis are community vs. pathway mixing are used to infer useful biological information.
Publications
The results of this thesis have been published in:
Refereed Journal Papers
• G. Bosque et al. (2014).
“Topology analysis and visualization of Potyvirus protein–protein interaction network”.
In: BMC Systems Biology 8.1, p. 129
• A. Folch-Fortuny et al. (2016).
“Fusion of genomic, proteomic and phenotypic data: the case of potyviruses”.
In: Molecular BioSystems 12.1, pp. 253–261
• Y. Morales et al. (2016).
“PFA toolbox: a MATLAB tool for Metabolic Flux Analysis”.
In: BMC Systems Biology 10, p. 46
• M. Beguerisse-Díaz et al. (2016).
“Context-dependent metabolic networks”.
In: arXiv:1605.01639 [physics, q-bio]
Conference Presentation and Posters
• Bosque, G. and Picó, J. (2013).
Redes de Interacción de Proteínas. XI Simposio CEA de Ingeniería de Control.
Automática y Biología celular: unacombinación emergente.
• Vignoni, A., Bosque, G., Tabor, J., and Picó, J. (2013).
How to tell bistable cells in which state they should be? On modeling of population fraction control using light.
6th International Meeting on Synthetic Biology (SB6.0).
• Bosque, G., Picó, J., Folch-Fortuny, A., Ferrer, A., and Elena, S. (2014).
Topological analysis and visualization of Potyvirus protein-protein interaction network.
Advanced Lecture Course on Systems Biology.
• Bosque, G., Picó, J., Folch-Fortuny, A., Ferrer, A., and Elena, S. (2014).
Latent Structures-based Modeling of Mutated Protein-Protein Interaction Networks.
12th International Conference on Computational Methods in Systems
Biology (CMSB 2014).
• Bosque, G., Picó, J., Folch–Fortuny, A., Ferrer, A., and Elena, S. (2015).
Genomic, proteomic and phenotypic data fusion in potyviruses.
III Reunión de la Red Española Interdisciplinar de Biofísica de Virus (BioFiViNet 3).
Physical Virology: From Structure to Evolution.
• Bosque, G., Picó, J., Beguerisse-Díaz, M., Oyarzún, D., and Barahona, M.(2015).
Community detection in E. coli metabolic network using Markov Stability and constrain–based modeling.
2nd Symposium on Complex Biodynamics and Networks.