Pedro Mendes, Ph.D. - US grants

Medicago truncatula is a close relative of the world's most important forage legume, alfalfa (Medicago sativa). It is a rich source of natural products, such as flavonoids, isoflavonoids and triterpenes, which impact its properties as a forage legume. The main experimental approach of this project is to perturb the expression of these natural products, and other areas of metabolism, by exposing cell cultures to biotic and abiotic elicitors. Use of cell suspension cultures will allow sufficient material to be collected and analyzed in parallel. Three experimental conditions have been chosen that mimic natural environmental challenges. The ultimate goal of this project is to generate a truly functional genomics data set for control and elicited cell cultures. Such data will encompass expressed sequence information and the associated mRNA, protein and metabolite identities and concentrations. This project will produce a variety of data so it becomes imperative to establish integrative models and software to facilitate relational analysis of the data to each other and to previous knowledge on sequences and pathways. Software is a facilitator of the discovery process when it enables the user to "navigate" the biological data in a dynamic and transparent way, requiring only the most basic computational skills. The bioinformatics component of this project will: i) construct a relational database to store all data; ii) construct an expandable analysis server that will facilitate processing the data with several statistical and numerical algorithms; and iii) integrate the above components through a web interface. The data as well as the software will be made available publicly.
The data generated by this project will be used to construct a quantitative predictive model of the time courses after elicitation, which is required to interpret the regulation of the underlying complex biological processes. The data will provide information about the extent and nature of gene expression reprogramming in response to biotic and abiotic signals at the transcription, translation and metabolic levels. There will also be practical applications in directed gene discovery for important agronomic traits involving plant natural products. Finally, this project will make available to the scientific community a bioinformatics system capable of supporting functional genomics ranging from the transcriptome to the metabolome.

Humans, unlike most animals, cannot synthesize vitamin C and thus have an absolute dietary requirement for this nutrient. Plants are the primary source of dietary vitamin C, which is key to several metabolic functions including collagen synthesis. Because vitamin C is a good anti-oxidant it can also be used as a natural preservative. Fresh produce with higher levels of vitamin C should therefore have improved shelf-life and consumer appeal. Additionally, increasing vitamin C in plants would result in higher intake per portion of fresh fruit or vegetables that would have a positive impact on human health. Toward this goal we will engineer the model plant Arabidopsis thaliana for higher steady state levels of vitamin C. Arabidopsis, while not part of human diet, is an essential model because it is the only plant with a fully sequenced genome. Unlike animals, the vitamin C biosynthetic metabolic network in plants is at present still not completely known. A pathway for plant systems has been proposed however, a parallel pathway, possibly similar to the animal pathway, may also operate in plants. Evidence for this pathway comes from experiments in which a cDNA encoding L-gulono-g-lactone oxidase (GLOase), the terminal enzyme in the animal pathway was expressed in tobacco and lettuce, which resulted in a 5-7 fold increase in vitamin C levels. Specific aims of the project include metabolically engineering normal and vitamin C-deficient Arabidopsis mutants to express GLOase to determine if any of the plant pathway intermediates directly supply GLOase or its precursors. Genes encoding vitamin C pathway(s) enzymes will be identified in activation tagged Arabidopsis lines by ozone screening and vitamin C analysis to fill in remaining gaps in the pathway(s). The outcomes of this project are anticipated to result in (a) Arabidopsis transformants with larger pools of vitamin C than the wild type, (b) identification of plant genes related to metabolic steps of vitamin C synthesis, and (c) further knowledge of which steps of the animal pathway also operate in plants.

The long range goal of this project is to build methods to infer genetic networks from gene chip or microarray transcription profiles. The methods developed are also envisioned to enable experimental design. The approach proposed is a staged one consisting of initially building a low-resolution dynamic model that fits observed data, and then to improve such a model to identify hidden variables in order to increase its predictive power. To achieve that objective, a series of kinetic functions for gene expression will be developed and made available on the Internet (where they will be peer-reviewed). These models will be used to generate artificial gene expression data. The data analysis algorithms will thus be developed and tested using a known in silico source to assess robustness and areas for further improvement.

Abiotic stress in the form of drought, salinity, and cold has a major impact on grape production and quality. Several studies have shown that water-deficit-stressed grapevines produce superior quality wine. The molecular genetic and biochemical basis for this correlation remains poorly understood. An integrative and quantitative analysis of mRNA, protein, and metabolite changes following abiotic stress imposition is required to enhance production efficiency under stress conditions and to understand the plant-derived contribution to constituents of wine quality. It will also require the customization and application of comprehensive bioinformatics systems to track and analyze changes that arise in response to abiotic stress and potentially related aroma, flavor, and color characteristics of grape juice and wine. One long-term goal of our research is to develop comprehensive genomic tools to facilitate the genetic engineering of improved abiotic stress tolerance traits in V. vinifera. The specific objectives for accomplishing this goal include: 1) extensive gene discovery through large-scale expressed sequence tag (EST) sequencing and mRNA expression profiling using oligonucleotide microarray-based expression monitoring in roots, leaves, and fruits of grapevines exposed to multiple abiotic stresses; 2) global mRNA expression profile data will be complemented by protein expression analyses using state-of-the-art proteomics methodologies; and 3) identification of specific metabolites and metabolite profiles in grapevines and fruit following abiotic stress that confer desirable aroma, flavor and color quality characteristics and improved health benefits. Metabolite profiles from grape juice of well-watered and water-deficit-treated vines will be compared with quantitative data from mRNA and protein expression patterns using comprehensive bioinformatics systems to store and analyze data sets. Ultimately, these data sets will be integrated into a reliable prediction model for wine characteristics. The proposed research will greatly facilitate future gene discovery and enable improvements to be made in both production efficiency and wine quality under environmentally adverse growing conditions.

[unreadable] DESCRIPTION (provided by applicant): Modeling and simulation of biochemical networks has become an essential activity to aid in the understanding of cellular behavior and to facilitate quantitative interpretation of modern experiments. A new approach, "systems biology", is being advocated which combines modeling, simulation and quantitative experiments. The novelty of systems biology is that experiments, mathematical modeling, and computer simulation are practiced in a cooperative way, the results of one being used for subsequent iterations of the others. Arguably, to fulfill the promise of systems biology there is a need for software tools to be able to appropriately connect models to experiments. We propose to address this, and other systems biology needs, with continuing development of our software COPASI, which is the successor of the widely used Gepasi. This project will also provide support to the vibrant community of COPASI users/biomedical researchers. We will address this with the following Specific Aims: Aim 1: Enable COPASI with new functionality to provide advanced model analyses. This will include adding algorithms for nonlinear dynamical analysis, model complexity reduction, improved parameter estimation and methods for closing the loop of experiment-model-experiment, and parallelization of certain algorithms. Aim 2. Improve and extend interoperability and standards compliance. COPASI will be equipped to facilitate users to create and maintain models according to the MIRIAM standard, and with methods to act as a client and provider in the SBW environment. SBML support will continue and be enhanced. Aim 3. Software maintenance. A testing plan will be established, including appropriate testing suites and procedures; bug reports will be collected and fixed; suggestions for improvement will be collected from users and followed. Aim 4. Support the modeling community. Providing online support tools, such as a user forum, FAQ and Wiki. An Annual COPASI User's Workshop will be held. [unreadable] [unreadable] [unreadable] [unreadable]

PROJECT SUMMARY  Biomedical  research  is  becoming  increasingly  dependent  on  construction  and  simulation  of  computational  models. Arguably this will be even more the case with the development of personalized medicine. However,  the technical aspects of modeling and simulation are overwhelming to many biomedical researchers. What is  needed is a software application capable of providing the appropriate numerical algorithms, but shielded by a  user  interface  that  aides  the  biomedical  researcher  in  conducting  the  required  simulations.  Modeling  and  simulation  can  be  applied  at  the  level  of  molecules,  their  networks,  cells,  tissues  and  whole  organisms.  Sometimes several of these levels have to be represented in order to properly predict and understand health  and  disease.  Thus  models  are  becoming  larger,  multiscale,  and  require  various  different  mathematical  frameworks  for  simulation.  We  propose  to  address  this  need  with  continuing  development  of  the  COPASI  software, which is already widely used in the biomedical research community, addressing the current trends.  This project will also provide support to the vibrant community of COPASI users/biomedical researchers. We  will address this with the following Specific Aims:  Aim 1. Add new numerical simulation and analysis methods to further support biomedical research. We will  develop  and  add  new  hybrid  simulation  algorithms  to  address  models  that  require  some  of  its  parts  to  be  simulated in different frameworks. We will add a new task to analyze parameter identifiability, which is very  useful for finding out if the model and data are matched, or improvements need to be made in both.  Aim 2. Improve COPASI?s user interface and the interfaces with other software. We will improve the graphical  user  interface  to  allow  it  to  efficiently  manipulate  very  large  models.  We  will  create  a  new  programming  interface so that others can easily use COPASI?s functions from other programs.   Aim 3. Software maintenance and standards compliance. We will continue to maintain the software, correcting  errors and making improvements, guided by feedback collected from its users. We will continue to implement  standards compliance to ensure the software is interoperable with other applications.  Aim  4.  Support  the  modeling  community.  We  will  continue  outreach  program  activities  aimed  helping  biomedical researchers make full use of the software?s capabilities. This includes continuing to offer tutorials,  courses, and workshops? to create further training videos and to maintain a web­based discussion group.    

PROJECT SUMMARY The project addresses an important biomedical problem: how to control biofilms formed by Candida albicans, a dimorphic fungus that is an important cause of both topical and systemic fungal infection in humans, in particular immunocompromised patients. It is responsible for 85-95% of all vaginal infections resulting in doctor visits. C. albicans biofilms also form on the surface of implantable medical devices, and are a major cause of nosocomial infections. In recent years, it has been recognized that interactions with bacterial species integrated into biofilms can affect C. albicans virulence and other properties, It is therefore important to understand the interactions of C. albicans with bacterial species, in particular metabolic interactions. The next step then is to understand and, ultimately, control how varying compositions of the different microbial species affect their metabolic state and their ability to form biofilms. This project approaches the problem through model-based design of optimal compositions of the bacterial species for control of fungal growth. This will be accomplished through a combination of the construction of a novel computational model of a heterogeneous biofilm consisting of bacterial as well as fungal species, and novel mathematical tools for dimension reduction and optimization. The outcome of the project will be a better understanding of the relationship between bacterial and fungal species in a biofilm and its therapeutic potential through the construction of a predictive agent-based computational model. Another outcome will be a mathematical tool that enables the use of mathematical models for the purpose of designing optimal controls for fungal growth in heterogeneous biofilms. The applicability of the results of this project extends far beyond biofilms into all areas of medicine and healthcare that are amenable to agent-based modeling, such as studies of the human microbiome.

Project Summary This proposal aims to establish a National Resource for Mechanistic Modeling of Cellular Systems to serve the large community of cell and systems biologists. The Resource will encompass COPASI and Virtual Cell (VCell) software platforms, which are arguably the most comprehensive and widely used tools for computational modeling of the biophysical mechanisms controlling cell function. VCell supports a number of key biophysical mechanisms, including reaction kinetics, diffusion, flow, membrane transport, lateral membrane diffusion, electrophysiology and rule-based models of multi-state/multimolecular interactions. Simulations can be based on 0D, 1D, 2D or 3D analytical or experimental image-based geometries. Users may choose among multiple available simulation approaches: ordinary differential equations, partial differential equations, stochastic reaction kinetics, network-free simulations, spatial particle-based simulations and spatial hybrid stochastic/deterministic simulations. COPASI enables the simulation and analysis of complex biochemical reaction networks either deterministically or stochastically. It offers a broad range of analysis tools including parameter estimation/optimization, steady state analysis, stoichiometric analysis, sensitivity analysis and metabolic control analysis. VCell and COPASI each boast thousands of active users. Collectively, their users produce >100 papers per year relying on these software systems. Both VCell and COPASI will be hosted at the University of Connecticut School of Medicine. Thus, the Resource will benefit from: (1) the common institutional organization under which it will operate; (2) a joint website; (3) a common high performance computing facility that will serve the computationally intensive needs of users; (4) coordinated training and outreach to the user community in the form of web-based documentation and tutorials, a yearly Computational Cell Biology Workshop and numerous roadshows at national and international meetings. Additionally, VCell and COPASI will be leveraged by external systems biology software developers as user-friendly platforms for the wide dissemination of their third party algorithms, software and databases. Finally, the Resource will actively engage with the software standards community to assure the reproducibility and reusability of both the software and the models it generates.