U SING GENOMIC - BASED INFORMATION FOR THE MODELING OF BACTERIAL ENVIRONMENTS AND LIFESTYLE Omer Eilam Eytan Ruppin School of Computer Sciences Tel Aviv University January 2010
W HY STUDY METABOLISM ? 1. It’s the essence of life… 2. Tremendous importance in Medicine a. Metabolic diseases (obesity, diabetics) are major sources of morbidity and mortality. b. Metabolic enzymes and their regulators gradually becoming viable drug targets. 3. Bioengineering applications Design strains for production biological products a. of interest. Generation of bio- fuels. b. 4. Probably the best understood of all cellular networks: metabolic, PPI, regulatory, signaling
Species evolve to adapt to their environment .
Phenotype Genotype Environment/lifestyle Can we use the genotype to predict the lifestyle of a species?
F ROM GENOMIC INFORMATION TO PHENOTYPIC ( METABOLIC ) INFORMATION Gene Genomic Metabolic Enzyme information information Genomic Metabolite information Metabolic information Hundreds of fully sequenced bacterial species
N EW APPROACHES ALLOW RECONSTRUCTION OF SPECIES ’ METABOLIC - ENVIRONMENTS Based on the network topology, identifying the set of compounds that are exogenously acquired External metabolite Internal metabolite From Borenstein et al, PNAS 2008
C ONSTRUCTING PREDICTED ENVIRONMENTS ACROSS HUNDREDS OF SPECIES Internal Metabolite External Metabolite Metabolic Predicted metabolic information environments
C HECKING VIABILITY FOR EACH SPECIES ON EACH ENVIRONMENT Internal Metabolite External Metabolite Essential Biomass Component Metabolic information
W E NOW HAVE AN ENVIRONMENTAL MODEL Environmental viability matrix Viable Not viable Env 1 Env 2Env N Spc 1 Spc 2 Information on species Spc N Information on environment
W HAT IS IT GOOD FOR ? Species • Can we characterize the lifestyle of a species based on Metabolic Genomes Genomic attributes? networks (Freilich et al, Genome Biology 2009) • Can we characterize ecological ? strategies based on genomic attributes? (Freilich et al, Genome Biology 2009) • How does the structure of the Environments metabolic network reflect adaptation to species’ lifestyle? (Freilich et al, PLoS Comp Biol 2010) • Can we characterize ecological communities based on genomic attributes? (Freilich et al, Under revision)
F IRST QUESTION Species • Can we characterize the lifestyle of the species based on Metabolic Genomes Genomic attributes? networks Can we predict, based on genomic knowledge, whether a species ? is a specialist or generalist? Can we estimate the range of environments it can inhabit? Environments
G ENOMIC - BASE PREDICTED DIVERSITY CORRESPONDS WITH ECOLOGICAL KNOWLEDGE Available systematic Genomic- based predicted estimates/information for environments environmental variability NCBI Fraction of reg. Specific examples : annotations genes Pseudomonas √ √ √ aeruginosa Multiple High Desulfotalea x √ x psychrophila Low Specialized Beyond specific examples: strong correlations (>0.3) between the metabolic-environment variability and established measures of environmental variability
S ECOND QUESTION Species • Can we characterize complex ecological attributes based Metabolic Genomes on genomic attributes? networks Can we predict the level of competition a species encounters ? in its natural environments and its rate of growth? Environments
A PPLYING THE ENVIRONMENTAL - MODEL FOR THE CHARACTERIZATION OF ECOLOGICAL ATTRIBUTES – COMPETITION From Freilich et al, Genome Biology, 2009 Environmental Viability Matrix Env 1 Env 2 Env 3 Env 4 Viable Spc 1 Not viable Spc 2 Max-CHS Spc 3 Spc 4 Co-Habitation vector Max-CHS {1,3,2} 3 Spc 1 Lifestyle annotation Spc 2 {3} 3 Spc 3 {3,2} 3 Spc 4 {1} 1
D ELINEATING ECOLOGICAL STRATEGIES FOR RATE OF GROWTH : Ecological diversity with intense co-inhabitation, associated with a typical fast rate of growth. Maximal co-habitation Freilich et al, Genome Biology, 2009 A specialized niche with little co-inhabitation, associated with a Environmental diversity typical slow rate of growth
I NTERIM SUMMARY The patterns observed suggests Species Metabolic a universal principle where networks metabolic flexibility is associated Genomes with a need to grow fast, possibly in the face of competition. ? The interplay between the environmental diversity – and maximal co habitation allows training a predictor for growth Environments rate (ROC score of 0.75 ).
M ETABOLIC N ETWORK M ODELS The application of computational methods to predict the network behavior usually requires additional data other than the network topology A ‘metabolic network model’ is a collection of such data: Reaction stoichiometry Reaction directionality Cellular localization Transport and exchange reactions Gene-protein-reaction association
M ODEL R ECONSTRUCTION P ROCESS (I)
R ECONSTRUCTION OF E. COLI MODELS
A VAILABLE M ETABOLIC M ODELS
S TOICHIOMETRIC M ATRIX (II) Stoichiometric matrix – network topology with stoichiometry of biochemical reactions (denoted S)
K INETIC M ODELING : D EFINITION Predict changes in metabolite concentrations m – metabolite concentrations vector - mol/mg S – stoichiometric matrix v – reaction rates vector - mol/(mg*h) d m A set of Ordinary Differential S v S f ( m , k ) Equations (ODE) dt Reaction rate equation Kinetic parameters • Requires knowledge of m, f and k!
C ONSTRAINT - BASED MODELING (CBM) (I) • Assumes a quasi steady-state – No changes in metabolite concentrations (within the system) – Metabolite production and consumption rates are equal • Represents the ‘average’ flow in the network over a long enough period of time d m S v 0 dt • The reaction rate vector v is referred to as a ‘steady -state flux distribution’ • No need for information on metabolite concentrations, reaction rate equations, and kinetic parameters
CBM (II) • In most cases, S is underdetermined, and there exists a space of v S 0 possible flux distributions v that satisfy: • The idea in CBM is to employ a set of constraints to limit the space of possible solutions to those more likely/correct – Mass balance is enforced by the above equation – Thermodynamic: irreversibility of reactions – Enzymatic capacity: bounds on enzyme rates – Availability of nutrients Solution space Correct solutions
F LUX B ALANCE A NALYSIS (I) An optimization method for finding a feasible flux distribution that enables maximal growth rate of the organism Based on the assumption that evolution optimizes microbes growth rate To enable maximal growth rate the essential biomass precursors (metabolites) should be synthesized in the maximal rate Add to the model a pseudo ‘growth reaction’ representing the metabolites required for producing 1 g of the organism’s biomass These precursors are removed from the metabolic network in the corresponding ratios: 41.1 ATP + 18.2 NADH + 0.2 G6 P… -> biomass
F OR EXAMPLE : B IOMASS REACTION OF E. COLI
P REDICTING GROWTH RATE X axis – Succinate uptake rate Y axis – Oxygen uptake rate Z axis - Growth rate (maximal value of the objective function as a function of succinate and oxygen uptake)
D OES E. COLI BEHAVE ACCORDING TO THE MODEL PREDICTIONS ? Succinate/oxygen PPP The experimentally determined growth rates were similar to the ones predicted by the model Ibarra et al., Nature 2002
P REDICTING K NOCKOUT L ETHALITY (I) A gene knockout is simulated by setting the flux through the corresponding reaction to zero. The corresponding reactions are identified by evaluating the gene-to-reaction mapping in the model.
G ENE KNOCKOUT LETHALITY : E. COLI IN GLYCEROL MINIMAL MEDIUM In total, 819 out of the 896 mutants (91%) showed growth behaviors in glycerol minimal medium in agreement with computational predictions.
M ETABOLIC MODELING OF A MUTUALISTIC MICROBIAL COMMUNITY Producing and analyzing the first multispecies stoichiometric metabolic model Prediction of several ecologically relevant characteristics Sulfate (Stolyar et a l, 2007)
A three compartments model: D. vulgaris metabolic model M. maripaludis metabolic model Culture medium
W HERE ARE WE HEADING ? The reconstruction process is continuously shortened, and species models are beginning to accumulate. This enables us to go large-scale and address fundamental ecological questions with the more powerful tools of constraint based modeling.
Shiri Freilich Elhanan Borenstein Adi Shabi Keren Yitzchak Tomer Shlomi Uri Gophna Roded Sharan Martin Kupiec Eytan Ruppin T HANKS
Recommend
More recommend
