资源描述
Software URLs:http:/cytoscape.org/Follow the online tutorial http:/www.pathogenomics.ca/cerebral/Follow the online tutorialhttp:/innatedb.ca/Follow the tutorials under HelpFollow the online tutorialhttp:/www.biotapestry.org/quickStart/QuickStart.html Follow this link for tutorial Try the models belowTry the models below.1: ODE modeling with Berkeley Madonna Install and run the freely available Berkeley Madonna (BM) software from , then copy the model below. You can download the Berkeley Madonna manual separately from When you first invoke Berkeley Madonna, it comes up with a registration dialog box. If you dont want to purchase the full version, simply click cancel, then click FILE NEW to open a new model file. In Berkeley Madonna, comment lines start with a semi-colon (;) and the prime symbol () can be used to denote the derivative of a variable (e.g. C is the same as dC/dt). In the free version of BM, you cannot save model files. But you can save your equations as plain text files (i.e. with extension .txt), which BM can read.Below is the Berkeley Madonna model description file for an athlete accelerating at the start of a race. The athletes acceleration is defined as the second derivative of her position (position”). position gives her speed. Plot the athletes position, speed and acceleration over time and verify that the area under the acceleration curve gives the speed and the area under the speed curve gives the position.METHOD RK4STARTTIME = 0STOPTIME= 25DT= 0.01acceleration= 0.3init position= 0init position= 0position= accelerationThe first line of the file specifies the numerical integration method (BM has several). The next three lines specify the simulation settings. DT is the step size for the numerical integrator (how often the ODE is evaluated). The model description starts by defining the parameters and initial conditions (i.e. at time zero, the start of the simulation). The last line defines the ODE.To change the variables plotted, double-click in the plot window. To change the axis settings, double-click on the axes.Here is the 2-ODE model of mRNA and protein levels. For simplicity, all the kinetic parameters set to one. A semicolon marks the start of a comment in BM.METHOD RK4STARTTIME = 0STOPTIME=10DT = 0.02Y = step(1,1); Y is defined as a step function going from 0 to 1 at time 1.init mRNA = 0init P = 0mRNA = Y - mRNAP= mRNA PIn addition to time-course simulations, you can use the “Parameter Plot” feature of BM (in the Parameters menu at the top) to plot instantaneous or steady state levels of mRNA and P as a function of Y. You can also use the “Batch Run” feature to plot the effect of parameter variations, or use “Define Sliders” to explore parameter changes interactively. Another very useful feature is the “Curve Fit” function, which allows you to fit parameter values to experimental data.For a summary of Berkeley Madonna syntax and functions, click HELPEQUATION HELP.2: The nonlinear behavior of a regulatory complex (uses Berkeley Madonna)The model below simulates mass-action kinetics of a biphasic complex formation process. To plot the steady state value of the complex as function of scaffold levels, use the Parameter Plot facility in Berkeley Madonna. Click in the plot window first, then click PARAMETERS PARAMETER PLOT from the top menu. Choose 100 runs, initial S value 10, final S value 1000, series type geometric), and the final value of the complex concentration for the Y axis. Repeat for complex vs. P2 and compare the results.METHOD RK4; defines the ODE numerical integration methodSTARTTIME = 0STOPTIME= 10; length of simulation in arbitrary time unitsDT = 0.0001; step size of the ODE numerical integratorP1= 100 (s_P1 + complex); total P1 =100 (arbitrary) unitslimit P1= 0; just a precaution to avoid negative amountsP2= 100 (s_P2 + complex); total P2 =100 (arbitrary) unitslimit P2= 0S= 100 ; total scaffold =100 (arbitrary) unitsscaffold= S (s_P1 + s_P2 + complex)limit scaffold= 0init s_P1= 0; partial complex formations_P1= P1*scaffold - s_P1init s_P2= 0s_P2= P2*scaffold - s_P2init complex= 0; full complex formationcomplex= scaffold*P1*P2 + P2*s_P1 + P1*s_P2 - complexThe model below integrates the above model with the earlier 2-equation model of transcription and translation. Note that I have set Y = complex2 / (Kdiss2 + complex2) to mimic the case where the complex has 2 binding sites that cooperatively activate transcription (note most parameters are arbitrarily set to 1 to make the model easier to follow).METHOD RK4; defines the ODE numerical integration methodSTARTTIME = 0STOPTIME= 10; length of simulation in arbitrary time unitsDT = 0.0001; step size of the ODE numerical integratorP1= 100 - (s_P1 + complex); total P1 =100 (arbitrary) unitslimit P1= 0; just a precaution to avoid negative amountsP2= 100 - (s_P2 + complex); total P2 =100 (arbitrary) unitslimit P2= 0S= 100 ; total scaffold =100 (arbitrary) unitsscaffold= S - (s_P1 + s_P2 + complex)limit scaffold= 0init s_P1= 0; partial complex formations_P1= P1*scaffold - s_P1init s_P2= 0s_P2= P2*scaffold - s_P2init complex= 0; full complex formationcomplex= scaffold*P1*P2 + P2*s_P1 + P1*s_P2 - complexKdiss= 50Y = complex2 / (Kdiss2 + complex2)init mRNA = 0init P = 0mRNA = Y - mRNAP= mRNA PIf you “parameter plot” the final levels of the complex and protein against the scaffold levels, you should see something like: (note logarithmic S axis)3: Stochastic simulation using DizzyBelow is a stochastic model of a Michaelis-Menten type enzymatic reaction. Substrate (S) molecules are converted into product (P) molecules by an enzyme (E). You can run this model by downloading and installing the Dizzy simulator from .In Dizzy files, the symbol / at the start of a line denotes a comment and is ignored by the simulator. The model definition syntax is straight forward. First, a model name is given (following the # symbol). Next, the initial number of molecules is specified for each molecular species. In Dizzy syntax, all initial and total amounts, and kinetic rates are defined simply asvariable or parameter name = variable or parameter value;Reactions are all elementary (i.e. single reaction steps involving one or two molecular species). Each reaction definition starts with a single-word name (e.g. make_product, underscores can be used to join up words). Finally, the reactions are specified in the usual biochemical notation, each followed by a stochastic reaction rate constant (c in the lecture notes). Note the punctuation marks, which are necessary components of Dizzy syntax.#model michaelis;/ This is a simple model for exploring Michaelis-Menten enzyme kinetics/ The model is part of the Dizzy tutorial by Stephen Ramsey, 2004/11/18E = 25;S = 50;P = 0;ES = 0;enzyme_substrate_combine, E + S - ES, 1.0;enzyme_substrate_separate, ES - E + S, 0.1;make_product, ES - E + P, 0.01;For further guidance on Dizzy, please refer to the Dizzy web site and accompanying Guide and Tutorials. To run a simulation, click TOOLS SIMULATE from the menu at the top. Then type the following values in the Dizzy simulation window (see figure below): simulation time=500, number of results points shown=100. Click select all in view symbols (top right), and the Gillespie-direct simulation engine (at left). To see the average behavior of the system for multiple simulation runs, increase the stochastic ensemble size. To run the same model in deterministic mass-action mode, select one of the ODE solvers listed to the left of the window (e.g. ODE-RK5-adaptive).Click start, to run the simulation. Example simulation results using Gillespies algorithm (right hand panel) and one of the deterministic ODE algorithms (left hand panel) are shown below. Note that I did not modify the model in any way for these two simulations. The only difference between the runs is the choice of the numerical evaluation method (i.e. the model was not changed in any way).NB: If you change the text of the simulation model in Dizzy, you need to click TOOLS RELOAD MODEL before re-running the simulation (else Dizzy will run the old model).4: Transcription factor occupancy on DNA exact stochastic model vs. average DNA occupancy-#model occupancy;A= 10;B= 10;DNA= 1;kfA= 0.001;kfB= 0.001;krA= 0.01;krB= 0.01;kfB2AB= 0.001;krAB2B= 0.01;kfA2AB= 0.001;krAB2A= 0.01;DNA_A= 0;DNA_B= 0;DNA_AB= 0;A_binds_DNA,DNA + A - DNA_A,kfA;DNA_A_dissociates,DNA_A-DNA + A,krA;B_binds_DNA,DNA + B - DNA_B,kfB;DNA_B_dissociates,DNA_B- DNA + B,krB;A_binds_DNA_B,DNA_B + A - DNA_AB,kfB2AB;DNA_AB_releasesA,DNA_AB-DNA_B + A,krAB2B;B_binds_DNA_A,DNA_A + B -DNA_AB,kfA2AB;DNA_AB_releasesB,DNA_AB- DNA_A + B,krAB2A;/ look at DNA_AB variations over say 100 time units points/ next, we specify the average occupancy in terms of the above parametersoccupancy= 0;occupancyCumulative= 0;exactCumulative= 0;KA= kfA/krA;KB= kfB/krB;KB2AB= kfB2AB/krAB2B;KA2AB= kfA2AB/krAB2A;occupancy= (KB*KB2AB*B*A)/(1 + KA*A + KB*B + KB*KB2AB*B*A ); / to compare the two methods, co-plot these 2 cumulative valuesaverage_occupancy, - occupancyCumulative, occupancy;full_simulation,- exactCumulative, DNA_AB/(DNA + DNA_A + DNA_B + DNA_AB); / Can you explain why the exactCumulative is slightly less than / occupancyCumulative? (hint: look at values close to time zero)-5: stochastic transcription and translation/= mRNA transcription =/ All parameter values in this model are arbitrary (for illustration)kt= 2;DNA_AB= 0.5;RNA_0= 0;BTA= 1;translocation_rate= 0.75;RNA_1= 0;mRNA= 0;RNA_off_BP= 0;/ BTA=1 indicates the basal transcription apparatus is free/ Dizzy has a multi-step reaction function (useful for tranlocation!). / The number after step indicates the number of steps.transcript_start,BTA - RNA_0,kt*BTA*DNA_AB;clear_BP,RNA_0- RNA_1,translocation_rate, steps: 10;/ Now we have cleared the basal promoter (promoter escape).transcribe1bp,RNA_1 - RNA_off_BP+BTA,translocation_rate;full_transcript,RNA_off_BP- mRNA,translocation_rate, steps: 100;/= protein synthesis =transportRate= 1.0;c_mRNA= 0;kdmrna= 0.001;move_to_cytosol, mRNA - c_mRNA,transportRate;mRNA_degradation,c_mRNA- ,kdmrna;ks= 2;ribosome= 1;AA_0= 0;AA_1= 0;AA_off_start= 0;protein= 0;AA_synth_rate= 0.75;translation_start,ribosome - AA_0,ks*c_mRNA*ribosome;clear_start,AA_0- AA_1,AA_synth_rate, steps: 3;translate1AA,AA_1 - AA_off_start + ribosome,AA_synth_rate;full_protein,AA_off_start- protein,AA_synth_rate, steps: 30;kdp= 0.001;protein_decay,protein- ,kdp;-N.B. For faster simulation, we can replace the step-wise translocation process with a single delayed reaction step (see Dizzy manual).Example model 6: Simulation of the distribution of gene expression levels in a population of genetically identical cellsIn the preceding example, we explored the stochastic behavior of a simple reaction. The GRN model presented here shows how we could explore stochastic single-cell behavior for large numbers of cells using Dizzy.This toy model is provided for learning, and exploration purposes. To speed up simulations of large numbers of cells, I have approximated the RNA transcription and protein translation processes with single step reactions, and assumed that transcription-factor-DNA interactions reach rapid equilibrium, thus allowing the use of promoter occupancy functions (see lecture notes). You can augment this model with the more detailed model fragments presented in the chapter.The GRN modeled is shown schematically at right. Ligand molecules L (yellow triangles) activate receptor molecules R (blue Y symbol). A fragment of activated receptor molecules (Rn) nuclearizes and activates Gene1. Gene2 is activated by a ubiquitous input (U) and repressed by Gene1.Model definitionParameter values and predefined functions (i.e. library elements) can be described in a separate text file. They are included in a new model with the command #include. Predefined functions (e.g. for fractional saturation functions, and transcription/translation models) are instantiated by simply calling the name of the function and specifying the input and output variable mappings./ Dizzy model by Hamid Bolouri, Nov 2007/ Explanatory comments follow the symbol: /#include paras.dizzy; / import parameter values/- cis-regulation functions -#define fracSat3Activators(krA , krB, krC, A , B, C, fracSat3)/ promoter occupancy function for Gene1 (3 activators A, B, C)A_bound= krA*A;B_bound= krB*B;C_bound= krC*C;AB_bound= krA*A*krB*B;AC_bound= krA*A*krC*C;BC_bound= krC*C*krB*B;ABC_bound= krA*A*krB*B*krC*C;fracSat3= (ABC_bound) / (1.0+A_bound+B_bound+C_bound+AB_bound+BC_bound+AC_bound+ABC_bound);#define fracSatActivatorRepressor(krA, krR , A, R, fracSat1R)/ Promoter occupancy function for Gene2 / (activated by “A”, repressed by “R”)A_bound= krA*A;R_bound= krR*R;fracSat1R= A_bound / (1.0 + R_bound + R_bound*R_bound + A_bound);#define geneA( fracsat, mRNA, ks_mRNA, halfLifeMRNA, P, ks_P, halfLifeProtein)make_mrna, - mRNA, fracsat *ks_mRNA ;degrade_mrna, mRNA - , 0.7/halfLifeMRNA;make_protein, $mRNA- P, ks_P; degrade_prot, P -,0.7/halfLifeProtein;#define geneR( fracsat, mRNA, ks_mRNA, halfLifeMRNA)/ Simple model of transcription and translationmake_mrna, - mRNA, fracsat *ks_mRNA ;degrade_mrna, mRNA - , 0.7/halfLifeMRNA;/ - Signal Transduction -R=k3/k4;LR=0;Rn=0; / R is the receptorR_synthesis,- R,k3; R_decay,R- ,k4;ligand_receptor,$L + R- LR,k5;/ NB ligand concentration is constant throughout the simulationLR_diss,LR- R,k6;LR_activation,LR- Rn,k7;Rn_decay,Rn- ,k8;/ B and C are constant inputs to Gene1B=k12/k13;C=k14/k15;mRNA1 = 0; protein1 = 0;makeB,- B,k12;decayB,B- ,k13;makeC,- C,k14;decayC,C- ,k15;/- Genes -#ref fracSat3Activators fracsat3A (krA, krB, krC, Rn, B, C, fracSat3);#ref geneA gene1 (fracSat3, mRNA1, halfLifeMRNA1, ks_mRNA1, protein1, ks_P1, halfLifeProtein1);mRNA2 = 0;protein2 = 0;/ U is the ubiquitous activator of Gene2makeU,- U,k16;decayU,U- ,k17;#ref fracSatActivatorRepressor fracsat2 (krA, krR , U, protein1, fracSatR);#ref geneR gene2 (fracSatR, mRNA2, halfLifeMRNA2, ks_mRNA2);To run this model, copy it from above and paste it into a plain text file called singleCell.dizzy.Dizzy allows the above model to be simulated deterministically with Ordinary Differential, and stochastically with a variety of algorithms. Example deterministic and stochastic simulation plots (showing mRNA1 levels) are presented below. Generating parameter valuesTo generate different parameter values for each cell to be simulated, we can use the formula:sample_value = nominal_value*(1 + max_var*(2*rand-1);where max_var specifies the extent of variability between the parameters of individual cells, and rand is a random variable between zero and 1. The above formula can be embedded in a Microsoft Excel file to generate multiple sets of parameter values. See the file parametersForDizzy.xls for an example implementation. Sheet 1 of the file generates sample parameter values. You can copy and paste the relevant columns into a second sheet, then copy the contents of sheet 2 into a plain text file (see sheet 2 and the example parameter file paras.dizzy. Be sure to save the parameters file into the same directory in which you saved the simulation model.Sample parameter setHere is an example parameter file. To run the above model without generating your own parameters, copy and paste this file into a plain text file called “paras.dizzy”.krA = 0.010643;krB = 0.010290;krC = 0.010636;krR = 0.005160;k3 = 0.968239;k4 = 0.001007;k5 = 1.045423;k6 = 0.000962;k7 = 1.067699;k8 = 0.001014;B = 988.976041;C = 1038.913448;L = 1024.262026;U = 1058.964216;k12 = 0.010914;k13 = 0.000010;k14 = 0.010760;k15 = 0.000009;k16 = 0.010959;k17 = 0.000010;halfLifeMRNA1 = 57.027952;halfLifeProtein1 = 129.017806;halfLifeMRNA2 = 62.847672;ks_mRNA1 = 4.636519;ks_P1 = 0.451176;ks_mRNA2 = 5.393898;Simulating N cellsTo run the model on large numbers of single-cells, we need to call Dizzy repeatedly (once for each cell), specifying a different parameter-set file for each run. Suppose we have named our Dizzy model simulation file model.cmdl, and our data files are called i.txt (where i=1.N). The following DOS batch file copies one parameter file at a time to a temporary file called paras.dizzy, then calls Dizzy to run model.cmdl with paras.dizzy and output the results into a Comma Separated Values (CSV) file (which can be imported into Excel for plotting and analysis) A similar script can be written for Unix (as in Linux and Apple MacIntosh operating systems). If you use Unix, you will already be familiar with scripting. ECHO off SET PATH=C:.Dizzybinfor %x in (12345-N) do copy C:.%x.txt C:.paras.dizzy & runmodel -modelFile C:.model.cmdl -numSamples 300 -ensembleSize 1 -outputFile C:.%x.CSV outputFormat CSV-excel -stopTime 3000 -simulator gibson-bruckWhere the should be filled in appropriately with directory paths specific to your computer, and - N should be replaced with a string of numbers corresponding to your parameter set file names. The number of sample points in the results file (numSamples), and the simulation duration (stopTime) are set to 300 and 3000 in the above example, and can be changed as desired.7: Logic simulationBelow is a simple Berkeley Madonna logic simulation example.METHOD RK4STARTTIME = 0STOPTIME= 10DT = 0.01init in1= 0; in1 mimics a logic inputin1= pulse(1, 1, 2) + pulse(-1, 2, 2)init in2= 0; in2 mimics a 2nd logic inputin2= pulse(1, 3, 5) + pulse(-1, 6, 5)T= 0.6; T is the on-off thresholdORgate = in1T OR in2T
展开阅读全文