Evolutionary Systems Biology
EvoSysBio Meeting |
Abstracts 3 days of exploring the frontiers of biology Update 23 July 2012
Prof. David F. Anderson This talk will present a general introduction to the stochastic simulation of models arising in the life sciences. These models are often continuous time Markov chains and have a very specific mathematical structure. This structure allows us to represent their
Dr. David J. Baumler No other family of microorganisms has had a greater impact on human health then the Enterobacteriaceae, and these bacteria have evolved into a wide variety of commensal and human, plant, and avian pathogens. These organisms have diverged from a common ancestor ~300-500 million years ago (MYA), and little is known about ancestral metabolism. Using a paleo systems biology approach the metabolism of ancient microorganisms has been investigated through construction of metabolic models using either
Using these ancestral metabolic models I have analyzed the metabolic capacity for carbon, nitrogen, phosphorous, sulfur, and iron utilization in aerobic and anaerobic conditions and have identified conserved and differentiating catabolic phenotypes and found that in silico predictions accurately predict substrate utilization phenotypes to >80% accuracy when compared to experimental data. These findings lend new insight to how some of these metabolic strategies are used in numerous human niche locations where modern-day enterobacteria cause disease.
Bryan Biehl Genome-scale metabolic models are becoming powerful tools for investigating the nutrient requirements of an organism within an environment. To date, genome-scale metabolic models have been constructed for commensal and disease-causing enterobacteria, while no models exist for plant-pathogenic enterobacteria. We have generated the first genome-scale species-specific metabolic models for four genera of enterobacterial plant pathogens using comparative genomics to aid in model development. Each metabolic model contains over 800 genes and over 2000 reactions. Of these four plant pathogens, two of the organisms have a narrow host range (non-soft rot pathogens: Pantoea stewartii and Erwinia amylovora) and two have a broad host range (soft-rot pathogens: Pectobacterium atrosepticum and Dickeya dadantii). Each broad host range organism is capable of infecting more than 35 susceptible plants species whereas the narrow host range organisms are limited to one or two host species. We experimentally determined substrate utilization and compared these results to in silico predictions to validate and refine the models. Additionally, we are also able to predict growth for each plant pathogen with simulated conditions encountered in specific plant hosts. This work is the first to examine genome-scale metabolism of plant pathogenic bacteria and to identify differentiating metabolic strategies for each organism, offering new insight into their host ranges.
Prof. Karl Broman In a mouse intercross with more than 500 animals and genome-wide gene expression data on six tissues, we identified a high proportion of sample mix-ups in the genotype data, on the order of 15%. Local eQTL (genetic loci influencing gene expression) with extremely large effect may be used to form a classifier for predicting an individual's eQTL genotype from its gene expression value. By considering multiple eQTL and their related transcripts, we identified numerous individuals whose predicted eQTL genotypes (based on their expression data) did not match their observed genotypes, and then went on to identify other individuals whose genotypes did match the predicted eQTL genotypes. The concordance of predictions across six tissues indicated that the problem was due to mix-ups in the genotypes. Consideration of the plate positions of the samples indicated a number of off-by-one and off-by-two errors, likely the result of pipetting errors. Such sample mix-ups can be a problem in any genetic study. As we show, eQTL data allow us to identify, and even correct, such problems.
Prof. Gheorge Craciun Mathematical models are often used to analyze biological interaction networks, such as the dynamics of species in an ecosystem, the spread of infectious diseases within a population, and the dynamics of concentrations in biochemical reaction networks. Persistence and permanence are properties of mathematical models that provide information about the long-term behavior of the system. For example, the persistence property is relevant in deciding if, in the long term, a species in an ecosystem will become extinct, an infection will die off, or a chemical species will be completely consumed by a reaction network. We discuss criteria for persistence and permanence that are based on interaction network structure, and do not rely on the values of the parameters present in the model.
Prof. Michael C. Ferris While optimization is prevalent in many application areas, the use of modeling systems such as GAMS and AMPL have until recently been somewhat limited. We discuss what these modeling languages provide, how to use them in simple cases, what they lack, and a view on their architecture and possible extensions. Several examples will be given, along with some suggestions for design considerations in developing new languages or features.
Prof. Chris Hittinger Clear models show how selection can maintain variation at a single locus, but multiple epistatic loci, population structure, and asexual reproduction introduce potentially interesting complications. Extreme cases of balancing selection on single loci, such as at the major histocompatibilitycomplex (MHC), have resulted in trans-specific polymorphisms being maintained for millions of years. Surprisingly, we recently discovered a case where selection has maintained a 7-gene network located on 5 chromosomes in 2 distinct states (functional and not) in a species of yeast for millions of years. Experiments show that the structure of the gene network creates a bi-stable fitness landscape with partial gene networks being unfit. Sequence analyses suggest that this bi-stability led to selection for the gene network in some populations and against it in others, despite gene flow throughout the rest of the genome. Species with limited sexual reproduction and outcrossing, including many eukaryotic parasites like Plasmodium, are predicted to be especially prone to maintaining this kind of variation. Modeling this process would require parameters that include migration rates, sexual reproduction rates, outcrossing rates, mutation rates, and the conditional fitness of all possible genotypes. Some of these parameters can only be measured with limited precision and substantial investment, but sensitivity analysis might help put some realistic boundaries on the phenomenon and help predict how widespread it should be.
Justin S. Hogg Biological systems are most commonly modeled within the reaction network formalism, which follows system dynamics at the resolution of molecular species. Biological systems, however, are governed by local events at the scale of macromolecular domains. The multi-domain structure of macromolecules, combined with the local nature of interactions, may induce a combinatorial explosion that makes standard methods for handling reaction networks infeasible. Rule-based modeling (RBM) languages naturally capture the domain-based structure of local interactions found in biological systems. Molecular complexes are represented by hierarchical graphs: macromolecular domains are represented as vertices, macromolecules as groupings of vertices, and molecular bonding as edges. Reaction rules, which represent reactant motifs as subgraphs, govern local transformations of molecular graphs---including binding, site modification, degradation and synthesis. The formal nature of RBM languages permits simulation under a variety of semantics, including network-based ODEs, SSA, and network-free stochastic. Additionally, RBMs are amenable to formal analysis and model-reduction techniques. In this talk, I will introduce RBM in the context of the BioNetGen language. I will review the basic simulation methods and then present some recent developments, including compartmental modeling and hybrid approaches.
Prof. David Krakauer There is a tension between complex simulations and simple elegant equations that are sometimes found to describe fundamental physical laws. This tension is discussed. References:
Prof. Allan Laughon Animal development relies on secreted signals to pattern tissues and control growth. Secreted bone morphogenetic proteins (BMPs) have diverse roles, including specification of the dorsal-ventral axis and patterning of the neural tube, limb, lung and other organs. BMP signaling is conserved throughout the animal kingdom and its specific role in D/V and neurectodermal patterning in insects is homologous to that in mammals. Consequently, Drosophila has been a important model system for discovering how BMP signaling functions in a variety of contexts - e.g., different scales of distance and time - and to understand how signaling is transduced to regulate gene expression in responding cells. From this work we know that BMP gradients instruct cells as to their position within a developmental field, but that BMPs can also stimulate growth. But there is no clear answer as to how a BMP gradient can promote growth evenly across a developmental field while simultaneously specifying different cell fates based upon position along the gradient. Similarly, there currently is no generally accepted model for how patterning scales in concert with tissue growth. Multiple factors modulate BMP signaling, both outside and inside cells, some as negative feedback loops that impart elasticity. A feed forward mechanism specific to insects may serve to delay regulatory responses. Systems modeling of BMP signaling has been developed by several groups as a means of evaluating parameters that shape BMP gradients and determine response thresholds.
2:15 pm Prof. Miron Livny Some thoughts and reflections about the modeling and simulation of Discrete Event Systems. The discussion will be based on more than a decade experience in using such systems to study the performance of a wide range of computing systems.
Prof. Laurence Loewe Evolutionary Systems Biology necessarily touches many different disciplines. Does that mean anything could be defined as contributing to EvoSysBio? Ideas will be presented that grew out of discussions that have been going on for quite some time.
Prof. Laurence Loewe To build rigorous quantitative models of biochemical and other biological reaction networks is at the heart of many efforts in evolutionary (and other) systems biology. A substantial amount of effort is required by any serious modeler who wants to link the model to known data. The load of such work can be lighthened by an appropriately designed model description language. Evolvix is currently being designed as one such language that shall facilitate the implementation and analysis of quantiative models of reaction networks. A brief overview of Evolvix will be presented.
Dr. Philip Poon Actions that are executed with some delay play an important role in building realistic models of biological systems in a variety of applications such as gene regulatory networks, cell differentiation, epidemiology and ecology. A number of delay stochastic simulation algorithms have been published in recent years to help simulate efficiently processes where both, rare events and delays play an important role. Using the simple model of a rabbit population that includes pregnancies and foxes that prey on rabbits, we demonstrate some of the pitfalls of existing algorithms and propose a solution that avoids granting pregnant rabbits accidental immortality.
Prof. Tom Rutherford This talk will describe programming frameworks for analyzing integrated economic-ecosystem models in which agent and species choices are characterized by optimizing behavior. In the ecosystem model all biomass transfers are mediated through energy prices. The General Equilibrium Ecosystem Model (GEEM) framework, originally developed by Tschirhart, addresses the complex interdependencies of ecosystems through the application of economic concepts to characterize the individual behavior of plants and animals. Each plant is assumed to maximize the difference between incoming solar radiation and energy lost to respiration and other species. For example, plants require access to light for photo-synthesis and obtain access by producing biomass that is exposed to the light. Greater biomass means more exposure to light but it also means more energy lost to the atmosphere in respiration; and therefore there is a limit to how much biomass a plant can produce. These models further accommodate explicit integration of ecosystem responses within an economic market equilibrium framework. In this talk, applications of the GEEM methodology to resource management issues will be surveyed and the use of an algebraic modeling language (GAMS) for the formulation, solution and analysis of these models will be discussed.
Prof. Garret Suen Ruminants like cows play an essential role in human agriculture. As dominant herbivores in North America, ruminants rely on microbial symbionts they harbour in their rumen to ferment plant biomass into nutrients for their host. Our own research is focused on characterizing and understanding the complex microbial community found within the rumen. We have been using a combination of whole-genome sequencing and RNA-seq to characterize the potential functional roles of specific bacteria within the rumen. Whole-genome sequencing of two of these cellulolytic bacteria, Fibrobacter succinogenes S85 and Rumincococus albus 7, reveal uniquely different approaches to cellulose degradation. Moreover, transcript sequencing of R. albus 7 on cellulose and cellobiose surprisingly revealed that the most highly expressed genes include the tryptophan biosynthesis operon. This has led us to determine that tryptophan is found in greater abundance in proteins associated with cellulose degradation, when compared to other genes in the genome. By applying these systems-level approaches we are beginning to understand how specific members of the ruminal bacterial community are contributing to the conversion of feed into host-usable nutrients.
Dr. Leslie Turner (1,2) Hybrid dysfunction, a common feature of reproductive barriers between species, is often caused by negative epistasis between loci (Dobzhansky-Muller incompatibilities). Identifying both (or multiple) interaction partners is challenging using traditional genetic approaches, hence little is known about the nature and complexity of hybrid incompatibilities. Male hybrids between recently diverged subspecies of house mice (Mus musculus) often show reduced fertility. We used a systems genetics approach to identify disruptions in gene networks associated with sterility. We collected genome-wide testis expression data from 305 male F2s from a cross between wild-derived inbred strains of M. musculus musculus and M. m. domesticus. We identified several thousand cis- and trans-acting quantitative trait loci (eQTL) contributing to expression variation.
Prof. Douglas B. Weibel Proteus mirabilis is a Gram-negative bacterium that typically resides in the soil. As an opportunistic pathogen, P. mirabilis also invades the urinary tract, often via the placement of urinary catheters. Cells of Proteus mirabilis swarm on a variety of different surfaces and produce a characteristic community structure over length scales that can approach tens of centimeters. Although reaction diffusion models can recapitulate patterns in silico that are similar to swarming colonies of P. mirabilis, we currently lack insight into biochemical and biophysical mechanisms that coordinate cell behavior and control the organization of communities. In this presentation I discuss quantitative studies of how flagella influence cell motility in physical environments that mimic swarming niches, such as viscous fluids.
Prof. John Yin The genome of every organism defines a process, and virus genomes are no exception. In an appropriate environment of a living host cell the release of a genome from an invading virus can take command, directing cellular material and energy resources toward the synthesis of components that are essential for virus growth: viral mRNA, viral proteins, and viral genomes. Assembly of these and other components yields progeny virus particles that, upon release by the cell, may then infect other susceptible cells. By performing quantitative experiments and building mathematical models of these processes we begin to link mechanistically how genomic information processing in limited resource environments can impact virus growth and infection spread. Here we will touch on three aspects of potential interest to evolutionary biologists: epistasis, population dynamics and niche construction.
|