Manual - Visual DSD
Contents
1. denotes concatenation along the upper strand This is a departure from the DSD syntax introduced in Phillips amp Cardelli 2009 and allows us to represent DNA molecules which are not formed by interactions along a common lower strand and permits reactions where a strand binds onto the upper strand of a gate This generalization seems necessary to allow DNA polymerization reactions The following example uses the visual notation to represent a complex DNA gate which uses both upper and lower strand concatenation gt An b t e f u V y DF Oe ae y 4 lt a gt b t gt lt c gt Slab et tu Vor etx tly lt z G gt W We impose a well formedness criterion on Visual DSD programs that no long domain d and its complement d can be simultaneously exposed in the initial program This suffices to ensure that no molecules can be produced which could interact on their long domains which preserves our key assumption that all reactions are toehold mediated Another important change in the latest version of Visual DSD is that strands and gates are considered to be equal up to rotation symmetry This is necessary because the distinction between upper and lower strands is simply an artifact of their representation on the page For example the following are considered to be equivalent b a a b C a b c lt c b a gt b E a c b a a b c e f z f e c b a af h cr er ar aa a b c lt d gt e f EE Cr eerie D7 22. 6 gt 5 INSUFFICIENT DNA SEQUENCES 3 If the Show Nucleotides checkbox is ticked the mapping of domains to nucleotide sequences is used to display those sequences in the graphical representations of molecules in the Initial tab and in the Species Reactions and Graph tabs When the checkbox is clicked these tabs are automatically refreshed to reflect the new setting As an example the following image shows one of the reactions from the Catalytic gate example with the nucleotide view enabled l RS KN GA A a k w CCCAAAACAAAACSAASACAA TATTCC CCCTITTCTAAACTAAACAA GCTA s gt CCCAAAACAAAACAAAACAA TATTCC CCCTITTCTAAACTAAACAA CCCTITTCTAAACTAAACAA GCTA GGGTIITGTIITTGTITTGTT ATAAGG GGGAAAAGATTTGATTTGTT CGAT GGGTITIGTITIGIITIGIT ATAAGG GGGAAAAGATTTGATTTGTT CGAT Note that the mapping from domains to nucleotide sequences is primarily intended for demonstration purposes and does not currently affect the behaviour of systems in any way reactions still place at the level of domains not at the level of individual base pairs Using Visual DSD from the command line In addition to the Silverlight interface Visual DSD can be run from the command line The command line tool contains a subset of the features of the Silverlight version except for the graphical visualizations and the compilation of domains into DNA sequences The program is typechecked and compiled as in the Silverligh
3. R Note that when we rotate a gate the domains in the double stranded segments are complemented in the syntax This is because we assume that the domain which appears in the syntax is the domain on the upper side of the double strand as viewed on the page Species are the kinds of molecule which may be included in a process As the following grammar shows gates and strands are classed as species as are instances of module definitions A module instance may include a list of values as parameters the number and type of parameters depends on the definition of the module If the wrong number or type of parameters are supplied the tool will signal an error Species Name Name Values Gate UpperStrand LowerStrand 16 Values The language of values in Visual DSD is fairly straightforward Values stands for a non empty comma separated list of Value elements Value String Integer Char Float Name true false Value Value Value Value Value Value Value Value float of int Value int of float Value Value Values Value Value Values The base types String Integer Char Float and Name follow the lexical conventions described above and the keywords true and false denote the usual Boolean constants There are standard arithmetic operators for addition subtraction multiplication and division Both arguments must be of the same type either integer
4. visualization of the initial and current states of the simulation run respectively This includes a graphical visualization of each molecular species along with their populations The graphical representation of species can be controlled via the dropdown labelled Options in the top right corner Options Ez Install __ _ Show Domains Editor Solver local Ji 5 Complement view al i Condensed view z _ Nucleotides view Domain format x comp _ xi Always display overhangs at an angle x Colour toeholds _ States warning Warn above 100000 sim points t Draw states graph _ Draw reaction graph Branch positioning Rotation Here we find a set of radio buttons which toggles between the default Complement view visualization mode which is used throughout this document the Condensed view used in Phillips amp Cardelli 2009 and a Nucleotides view which renders domains as actual nucleotide sequences discussed in the Implementing domains as nucleotide sequences section below Only one of these modes can be selected at a time The Colour toeholds checkbox toggles whether toehold labels should be written in the same colour as their lines in the visualization this is enabled by default It is possible to obtain fairly fine grained control over the behaviour of the Visual DSD compiler programmatically via directives A graphical entry point for
5. L 2009 A programming language for composable DNA circuits Journal of the Royal Society Interface Qian L amp Winfree E 2009 A simple DNA gate motif for synthesizing large scale circuits In G A F C Simmel amp P Sosik Ed DNA 14 LNCS 5347 pp 70 89 Springer Verlag Zhang D Y amp Winfree E 2009 Control of DNA strand displacement kinetics using toehold exchange J Am Chem Soc 131 17303 17314 Zhang D Y Turberfield A J Yurke B amp Winfree E 2007 Engineering entropy driven reactions and networks catalyzed by DNA Science 318 1121 1125 25 Appendix Collected summary of the Visual DSD language Grammar The full grammar of Visual DSD programs is as follows Terminal symbols of the language are written in teletype font and non terminals are in italics Program Directives Declarations Process Directive Directives Plots Declaration Declarations Parameters Directives Process Declarations Process Process directive directive duration Float duration Float points Integer directive directive directive directive directive directive directive directive directive directive directive directive Directive sample Float sample Float Integer scale Float concentration CU time TU plot Plots leak Float tau Float migrate Float lengths Integer Integer tolerance Float toeholds Float Float Directive Directives
6. amp Cardelli 2009 and it is the default selection when the program begins e Finite molecules are identified up to branch migration Strand displacement toehold covering and toehold unbinding reactions have a finite rate tau which can be adjusted using a directive as described above Consecutive tau reactions are merged together into a single reaction which is also assigned the rate tau Toehold binding has a finite rate as usual This allows us to model cooperative displacement while satisfying the constraint that toehold unbinding and strand displacement have similar rates e Detailed molecules are not identified up to branch migration Strand displacement branch migration toehold covering toehold binding and toehold unbinding reactions all have finite rates The strand displacement and branch migration rates are calculated from the elementary migration rate and the assumed values for domain lengths which can all be set with directives as described above Moving from the Infinite through to the Detailed semantic model the number of reactions in the system increases dramatically and the computational cost of the compilation and simulation processes increases also The same is typically true when unproductive reactions are included The leak model The Leaks checkbox enables the simulation of certain kinds of interference between gates and strands These allow for a more realistic simulation of the actual behaviour of the system b
7. Process constant Process constant Value Process Value constant Process Species new Name Value Value Process new Name Process Processes Process Process Processes Name Name Values Gate UpperStrand LowerStrand Segment Segment Gate Segment Gate Toehold LowerStrand UpperStrand Double Double Overhangs Overhangs Double Overhangs Double Overhangs LowerStrand UpperStrand 27 String Integer Single character Floating point Name True False Addition Subtraction Multiplication Division Convert integer to float Convert float to integer Parenthesised value Single value Commaz separated values Repetition Constant population Constant repetition Repeated constant Species Restriction with rates Restriction default rates Parallel processes Single process Parallel processes Module instance Module instance Gate Upper strand Lower strand Single segment Lower strand concatenation Upper strand concatenation Lower toehold Lower strand segment Upper strand segment Double strand Double with right overhang s Double with left overhang s Double with both overhang s Lower overhang Upper overhang Double UpperStrand LowerStrand Domains Domain Toehold Sequence TU CU LowerStrand UpperStrand UpperStrand LowerStrand Domains lt Domains gt Domains Domain Toehold Domain Domains Toehold
8. String Gate Strand sum Plots sub Plot Plot diff Plot Plot div Plot Plot Plots Plots def Name Process def Name Parameters Process new Name Value Value new Name def Name Value Declaration Declaration Declarations Name Name Parameters 26 A program must contain a process to run and may optionally contain directives and or declarations End time for simulation with optional number of datapoints End time for simulation with optional number of datapoints Specify a scaling factor Specify concentration units Specify time units Which species to plot Rate of leak reactions Rate of tau reactions Elementary migration rate Default domain lengths Specify ODE solver tolerance Default toehold reaction rates Single directive Multiple directives Plot exact string match only Gate species to plot Strand species to plot Plot sum of populations Plot population subtraction Plot population difference Plot population ratio Plot multiple populations Module definition Module definition Global channel with rates Global channel default rates Value assignment Single declaration Multiple declarations Single parameter Commaz separated parameters String Integer Char Float Name true false Value Value Value Value Value Value Value Value float of int Value int of float Value Value Value Value Values Value
9. application and a text based command line tool In order to run the Silverlight version you must install the correct Silverlight plug in for your operating system and web browser from http www microsoft com silverlight Silverlight compatibility has been tested on Windows and Mac OS X under various browsers including Internet Explorer Firefox Safari and Chrome With Silverlight installed browse to http lepton research microsoft com webdna and the user interface should load within a few seconds In the top right corner of the interface there is an Install button which you can click to install a local copy of the Silverlight program this should work both on Windows and on Mac OS We will describe the use of the Visual DSD Silverlight interface in the next section Next to the Install button is a License button which brings up a copy of the Visual DSD license agreement Once the software is installed the Install button becomes an Update button which can be used to check for and install any newer releases of the software Note that when a new version of the software is released online you may need to reset your web browser to delete the old version from the cache before your browser will load the new version There is also a command line version of Visual DSD which is included in the source distribution as pre compiled binaries for Windows and Mac OS X 10 5 The command line version can be comp
10. to produce a float value 1 One or more digits followed by a decimal point followed by zero or more digits For example 3 141 2 One or more digits followed by an uppercase E or lowercase e followed by a plus or minus sign followed by one or more digits For example 3e 5 3 One or more digits followed by a decimal point followed by zero or more digits followed by an uppercase E or lowercase e followed by a plus or minus sign followed by one or more digits For example 1 4324e 2 e Char a single character enclosed by apostrophes The character itself can be anything except for an apostrophe or a backslash Comments Comments are opened with and closed with They may be nested Reserved keywords The following identifiers are reserved for use as keywords of the programming language directive sample plot leak tau migrate lengths def mew true false int of float float of int time concentration constant tolerance sum scale duration points toeholds 29
11. Domains Sequence Sequence Sequence Sequence Integer Name seconds s minutes m hours h molar M milimolar mM micromolar uM nanomolar nM picomolar pM femtomolar fM attomolar aM zeptomolar zM yoctomolar yM 28 Lower and upper overhangs Lower and upper overhangs Double strand Single upper strand Single lower strand Single domain Toehold domain Multiple domains Multiple domains DNA sequence Complemented sequence Toehold sequence Complemented toehold DNA sequence DNA sequence Wildcard Unit of time is seconds Unit of time is minutes Unit of time is hours 1 mol L 10 mol L 10 mol L 10 mol L 10 mol L 10 mol L 10 8 mol L 10 mol L 10 mol L Lexical conventions The language uses the following lexical conventions We write digit for a single character in the range 0 9 and alphanumeric for any character in the range A Z or a z e Integer anon empty sequence of digits e Name the first character of a name must either be alphanumeric or the underscore character _ This is followed by a possibly empty sequence of characters which may be alphanumeric digits underscores or apostrophes e String a possibly empty sequence of characters enclosed by quotation marks Any quotation marks appearing within the outer of quotation marks must be escaped by a preceding backslash e Float there are three different ways
12. Visual DSD User manual v0 14 beta Matthew R Lakin Rasmus L Petersen amp Andrew Phillips Disclaimer This version of the manual is for the beta release only Note that many features present in the beta release are not yet described in this manual To use these undocumented features the best starting point is to consult the built in examples Introduction Visual DSD is an implementation of a programming language for composable DNA circuits based on that described in Phillips amp Cardelli 2009 The language includes basic elements of sequence domains toeholds and branch migration and assumes that strands do not possess any secondary structure The Visual DSD tool compiles a collection of DNA molecules into a set of chemical reactions It also includes a stochastic simulator which computes a possible trajectory of the system and graphs the populations of species over time and a deterministic simulator which forms and solves an ODE representation of the dynamics of the system Furthermore the reachable state space of the system can be constructed as a continuous time Markov chain This manual assumes that the user is familiar with the basics of DNA strand displacement and in particular the terminology and notation from Phillips amp Cardelli 2009 The reader is referred to that paper for the technical details of the language semantics Installing Visual DSD Visual DSD is available in two forms a Silverlight based graphical
13. ak reaction default is 10 nM s e Tau directives set the rate of a tau reaction in the Finite semantics default is 0 1126 s e Migrate directives set the rate of branch migration across a single nucleotide default is 8000 s The branch migration rate for a domain of length L is given by r L where r is the single nucleotide migration rate set by the directive e Lengths directives set default values for the lengths of toeholds and long domains For the purpose of computing rate constants all long domains are assumed to have the same length which is set using the lengths directive The code directive lengths 5 15 assigns a length of 5nt to toeholds and 15nt to long domains the default values are 6nt for toeholds and 20nt for long domains The value provided for toeholds must be greater than that provided for specificity domains or the system will raise an error Currently the assigned default length for toehold domains is not used to calculate rates Instead the user may set toehold binding and unbinding rates directly on a per toehold basis Note that specific nucleotide sequences are not currently used when computing rate constants e Tolerance directives specify the tolerance parameter of the deterministic simulator which provides a tradeoff between computational cost and smoothness of the resulting solution The default value is 10 It is crucial to choose a tolerance value which reflects t
14. at there are three different ways to produce a float value 1 One or more digits followed by a decimal point followed by zero or more digits For example 3 141 2 One or more digits followed by an uppercase E or lowercase e followed by a plus or minus sign followed by one or more digits For example 3e 5 3 One or more digits followed by a decimal point followed by zero or more digits followed by an uppercase E or lowercase e followed by a plus or minus sign followed by one or more digits For example 1 4324e 2 e Char a single character enclosed by apostrophes The character itself can be anything except for an apostrophe or a backslash Comments Comments are opened with and closed with They may be nested Reserved keywords The following identifiers are reserved for use as keywords of the programming language directive sample plot leak tau migrate lengths det new true false ant Of float Iloat Of int time Comcentracion constant tolerance sum scale duration points toeholds Programs Programs written in the language may consist of three parts in the following order 1 Directives to the simulator and plotter 2 Declarations of values global domains and modules 3 A process to run which contains the species and their initial populations A program must contain a process and may contain directives and or declarations Program Directives Declarati
15. ber of moles per unit volume The unit of concentration can be set by the concentration directive For example directive concentration M sets the units of concentration to molar The default unit is nanomolar nM where 1nM 10 mol L In order to perform a stochastic simulation concentrations must be converted to numbers of individuals This can be achieved using the following equation n cV Na where n is the number of individuals c is the molar concentration V is the volume and Na is Avogadro s constant which denotes the number of individuals per mole of substance approximately 6 02214x1023 mol The function x denotes the rounding up of x to its nearest natural number Thus in order to convert a concentration into a number of individuals it is sufficient to multiply the concentration by a scale factor s V Na which denotes the number of individuals per unit concentration Essentially this corresponds to choosing a volume V such that the number of individuals is equal to s for one unit of concentration For example a scale factor of 50 corresponds to choosing a volume that is 50 times the volume occupied by a single individual The unit of the 20 scale factor are assumed to be the inverse of the unit of concentration and are given as nM by default Note that the conversion from concentrations to individuals is achieved using a scale factor s rather than specifying a volume V directly since it is difficult to choo
16. centrations or populations of certain species The species to plot can be specified in the program but our example gives no such directives in this case the default behaviour is to plot the populations of all species The chart window can be dragged using the mouse and zoomed in and out using the scroll wheel Simulation States Inference Verification Export Chart Table Bars Initial Current Probability Map Probabilities Load data Save data Print _ Resample x Legend Top Right y Names Edit palette Select plots v 6500 N d lt 45 gt 6000 1 lt 1 gt 2 lt 6 gt 3 4 5 N 2 34 4 5 lt 1 2 gt N lt 6 3 4 gt 0 1000 2000 3000 4000 5000 6000 7000 Along the top of the plot window is a collection of buttons which give more control over the plot In the dropdown labelled Select plots clicking on the check box for a particular species toggles the visibility of the relevant line in the plot There is also a check box to show or hide all plots The check box labelled Names makes the legend use the names of species as defined in the program It is on by default but our program defines no such names so it has been turned off in the above screen shot as the auto generated names are less informative than the string representations When the simulation terminates or is paused the Initial and Current tabs are populated with a
17. d We present a brief overview here see Phillips amp Cardelli 2009 for a more technical discussion of certain aspects but note that Visual DSD expands on the syntax from that paper 13 We start with the most basic elements and work our way up The language does not work at the level of individual nucleotides instead the fundamental building blocks of species in Visual DSD are DNA sequences A sequence represents some finite section of DNA comprising many nucleotides see below for a discussion on mapping domains to nucleotide sequences We assume that distinct sequences are sufficiently different that they do not interact and that they do not possess any secondary structure The grammar for sequences in Visual DSD is as follows Sequence Integer Name A Sequence is represented by a name although this may actually be a number for convenience Hence a species suchas lt 1 2 3 gt isvalid where 1 2 and 3 denote different DNA sequences The underscore is permitted so that wildcards may be included in plotting directives see above A Domain can either be a sequence or its Watson Crick complement obtained by reversing the directionality of the nucleotide sequence then swapping C lt gt G and T lt gt A throughout We denote the complement of a sequence using an asterisk so the grammar for domains is as follows Domain Sequence Sequence Toeholds are DNA domains which are sufficiently short that the
18. eactions as described above e declare domains enabling this option raises an error if there are undeclared domains in the program for debugging purposes e profile produces additional profiling output such as the timings of various stages of the compilation process and the numbers of reactions produced e step lt float gt specify the percentage reporting step for the simulator e simulate lt string gt if this is enabled the system is simulated after it has been compiled and the results of the simulation are written into a CSV file As in the Silverlight interface there are three possible choices for the simulator mode Stochastic Deterministic and JIT e semantics lt string gt this modifies the semantic model used for computing the set of reactions As in the Silverlight interface the choices are Infinite Default Finite and Detailed The default semantic model is Default The command line arguments are the same in the command line version compiled using F except that ocamlrun is not needed in the command line invocation 24 Bibliography Cardelli L 2008 On process rate semantics Theoretical Computer Science 391 3 190 215 Fehlberg E 1969 Low order classical Runge Kutta formulas with stepsize control NASA Technical Report R 315 Huang C Y F amp Ferrel J E 1996 Ultrasensitivity of the mitogen activated protein kinase cascade PNAS 93 10078 10083 Phillips A amp Cardelli
19. es the accuracy of the stochastic simulation but can make the simulation algorithm run considerably faster Note that the Unproductive checkbox is greyed out when Infinite is selected from the Compilation menu because it has no effect under that semantic model The Leaks checkbox toggles the inclusion of leak reactions which are a form of unwanted interference between strands and gates see the Leak model section below for a brief explanation The Declare domains checkbox allows the user to enable stricter syntax checking which produces an error if a domain is used without having first been declared using the new keyword This can be helpful when debugging programs as a mistyped domain can be hard to detect and can significantly alter the behaviour of the system Finally the Polymers checkbox enables additional reduction rules which allow DNA gates to bind together to form polymers of potentially unbounded length This is a new feature that is not present in the language described in Phillips amp Cardelli 2009 The following reaction is an example of DNA polymerization in which two gates bind together to form a stable complex b a t b w b 4 T p r Epro However it is also possible to write DSD programs involving polymerization where the full reaction graph is infinite For example the following program directive sample 10 0 100 10 a aE lal Lote aAa produces var
20. escription of the species and the reactions in the XML based Systems Biology Markup Language for more information on SBML see the website at http sbml org This output can be saved directly into an XML file for use with external tools which accept SBML input Simulation States Inference Verification Export Chaste PRISM PRISM DTMC Domains Reactions Summary LES SBML Matlab C Zoom Save Load 7xml version 1 0 encoding UTF 8 gt a lt sbml xmins http www sbmil org sbml level2 version4 jevel 2 version 4 gt model gt het nithefinitinna gt Visual DSD provides limited support for producing nucleotide sequences for in vivo implementation of the system described by the user The Domains sub tab of the DSD Table tab displays a mapping from domains to nucleotide sequences see the Implementing domains as nucleotide sequences section below for details In addition to compiling DNA molecules into sets of chemical reactions the Visual DSD tool includes a simulator which can produce stochastic or deterministic plots of a trajectory of the system of reactions The simulator is accessible via the Simulate button in the top row This greys out the Simulate button and the simulation will run until it reaches its end point or until no further reactions are possible in our example this is time t 7000 the first number in the line of code which begins directive duration An ongoin
21. g a common graphical notation from the DNA computing literature see Zhang Turberfield Yurke amp Winfree 2007 for example Note that the default is a different graphical notation from that used in Phillips amp Cardelli 2009 although that notation can be selected as an alternative see below The shown molecules include all of the species which could possibly be produced by reactions from the initial species presented in the input program To see just the initial species one can click to the DSD Table tab and then the Species sub tab The CRN Graph tab useful alternative way to view the reactions in the system as a reaction network For the catalytic gate example the reaction network looks as follows Each labelled node in the CRN Graph tab denotes a species The initial species are represented with a bold outline Each unlabelled node represents a reaction which may or may not be reversible with edges connected to reactant and product species For irreversible reactions edges with no arrows denote reactants while edges with hollow arrows denote products For reversible reactions hollow and solid arrow heads are used to distinguish between the products of the forward and reverse reactions respectively CRN Simulate States Infer Venty Stop Pause _ Step DSD Code DSD Table CRN Code CRN Table CRN Graph Directives Graph ID E 9 Pan _ Zoom _ Layout _ Edge Layout Spli
22. g simulation can be paused using the Pause button the simulation can then be restarted from the beginning using the ungreyed Simulate button or unpaused using the same button as for pausing which will have been relabeled Resume Running the simulator produces output in the sub tabs of the Simulation tab on the right hand side Simulation States Inference Verification Export Chart Table Bars Initial Current Probability Map Probabilities Load data Save data Print _ Resample Legend Top Right x Names Ed p300 The output of the simulator is given as species populations at discrete points in simulation time The number of sample time points is specified in the program the second number in the line beginning directive duration By default as in the catalytic gate example the simulator samples the populations of all of the species in the system after every single reaction It is possible programmatically to restrict this to a subset of the species as discussed in the next section The Table tab contains a tabular representation of the simulation data as species populations over time The data in this tab can be saved as a comma separated CSV or tab separated TSV file which can then be imported into a spreadsheet such as Microsoft Excel Alternatively the data can be copied and pasted directly into Excel as it stands The Chart tab produces a real time graph of the con
23. hain of the system and displays it in various ways for analysis and output see the State space analysis section below for more details Programming language syntax Having learned how to use the Visual DSD tool on an example program the next step is to write our own programs in the language The system accepts user programs in the editor pane on the left hand side the font size can be adjusted using the Zoom slider The ASCII syntax of Visual DSD is an extension of that introduced in Phillips amp Cardelli 2009 We will describe this in detail over the rest of this section the collected grammar is also presented as an Appendix for reference purposes In this section the terminal symbols of the language are written in teletype font and non terminals are in italics Lexical conventions The language uses the following lexical conventions We write digit for a single character in the range 0 9 and alphanumeric for any character in the range A Z or a z e Integer a non empty sequence of digits e Name the first character of a name must either be alphanumeric or the underscore character _ This is followed by a possibly empty sequence of characters which may be alphanumeric digits underscores or apostrophes e String a possibly empty sequence of characters enclosed by quotation marks Any quotation marks appearing within the outer quotation marks must be escaped by a preceding backslash e Flo
24. he populations and reaction rates of the system in question or the performance of the deterministic simulator may suffer Note that the tolerance is multiplied by the scale factor in an attempt to maintain a reasonable value with respect to the species populations e Toeholds directives specify the default binding and unbinding rates in that order for domains which are declared without explicit rates The default values are 3 0 107 nM s for the binding rate and 0 1126 s for the unbinding rate Declarations These introduce new module definitions globally defined domains and value assignments Declaration def Name Process def Name Parameters Process new Name Value Value new Name def Name Value The first two lines are for module declarations which are written with the def keyword A module is simply a parameterised process Here Parameters stands for a non empty comma separated list of Names which are the parameters of that particular module The grammar also permits a module to have an empty parameter list We will describe processes below The name and parameters of a module are bound within the body of that module and the name of the module is bound in the remainder of the program A new declaration declares a new domain This is global in the sense that the name of the domain is bound in the rest of the program Optionally the domain may be annotated with two floating point value
25. iled using the F or Objective Caml compilers and contains build scripts to automate this process under Windows Linux or Mac OS X The command line executable implements most of the functionality of the Silverlight version We describe the use of the Visual DSD command line interface at the end of this document Using Visual DSD under Silverlight The Visual DSD tool comes with a number of example systems implemented using DNA molecules These are accessible from the drop down menu labelled Examples inside the tab labelled DSD Code in the top left corner of the Silverlight user interface Table CRN Simulate States DSD Code DSD Table CRN Code CRN Table oc e al PS 6 Ga GD Examples Catalytic Science 2007 Catalytic Catalytic Directives Tutorial VEMDP 2014 Join Join Reporter Join Reporter 23 Join Directives Join Events Join Fit Join Fit Errors The Catalytic example is an implementation of the entropy driven catalytic gate from Zhang Turberfield Yurke amp Winfree 2007 The Lotka example is the Lotka Volterra predator prey oscillator The Mapk example models a mitogen activated protein kinase MAPK signaling cascade Huang amp Ferrel 1996 and the Migrations example serves to demonstrate the branch migration rate model Zhang amp Winfree 2009 Most of the other examples are related to the implementation of chemical kinetics using DNA We will use the catalytic gate as a runn
26. ing example Selecting this populates the Code tab in the left hand pane with the text of the example program we can use the Zoom slider to adjust the text size We see that an asterisk has appeared in the tab title this indicates that the model has changed but not yet been compiled DSD Code DSD Table CRN Code CRN Table CRN Graph A simple catalytic circuit based on Zhang Science 2007 directive duration 7000 0 points 1000 directive scale 500 0 new 3 4 2E 4 4 E 2 new 5 6 5E 4 4 0E 3 13 lt 2 3 4 gt 10 lt 4 5 18 lt 1 gt 2 lt 6 gt 3 4 5 The text of this program begins with a directive to the simulator telling it the duration of the simulation run and how many sample data points to use The next line specifies a scaling factor 2 which the system uses to automatically scale up from molar concentrations to populations of individuals for the stochastic simulation See the discussion of directives below for more details on this and other constants that can be specified within DSD programs The third and fourth lines declare two domains with specified binding and unbinding rates The final element of the program is a collection of DNA molecules with their respective concentrations The syntax of the Visual DSD language is discussed in more detail below Now that we have a program to run clicking on the CRN button in the top row performs the compilation
27. into chemical reactions The program is checked for errors in the syntax and also for type errors such as when an integer is used where a floating point number would be expected If an error is found it is reported in a message box together with the location of the offending section of the program like the following example error message Line 5 char 2 8 Type error found float but expected int population of a species OK If the program is error free then the compilation process proceeds as described in Phillips amp Cardelli 2009 This produces output in all the downstream tabs and the focus switches to the Reactions sub tab of the CRN Table tab This output persists until a modified program is compiled using the CRN button CRN Simulate States Infer Venfy _ Step DSD Code DSD Table CRN Code CRN Table CRN Graph Directives Parameters O Species 9 Reactions 4 Add Remove Show Graphics Show Rates _ Show Names Catalysts Reactants Rate Products vd Co vd Co 2 3 4 4 5 i0 00065i 2 3 4 5 N 70 004 _ Ames f 7 i 7z 2 5 4 2 3 4 10 00042 2 3 J 4 1 2 2 3 4 2 3 4 5 2 3 4 2 3 4N 5 __ 0 004 sory 2 3 a Ni0 00065 vd So 7 vd 2 3 4 5 i0 047 2 4 5 6 3 4 a ane 70 00042 gt a 2 3 4 5 2 3 4 5 The Species tab visualises the DNA molecules in the system usin
28. ints stays constant but the simulation time is extended or shortened the resulting plot will be less or more detailed If no such directives are provided the default behaviour is to run the simulation for 1000 time units and take 10 000 samples of the species populations Scale directives allow the stochastic simulator to scale up from molar concentrations to populations of individuals Concentrations are scaled by simply multiplying by the factor and the rates of binary reactions are modified following Section 4 2 of Cardelli 2008 Thus the user does not have to worry about the details involved in switching between continuous and discrete simulation methods see below The scale factor is 1 0 by default The scale factor also modifies the tolerance parameter of the deterministic simulator as described below See the Populations and concentrations section below for further details Concentration directives allow one to specify the assumed unit of concentrations It will be printed on the y axis of simulation plots when the simulator is run in deterministic mode It will also feature in the summary information in the Text tab The default concentration units are nanomolar nM Time directives allow one to specify the assumed unit of time It will be printed on the x axis of simulation plots The default time units are seconds s Plot directives tell the system which populations to sample at each time poin
29. ious reactions including the following Clearly this DNA polymer could grow indefinitely if we kept supplying it with monomers to join onto the chain D x x x a b a a b r a b t a b A a c a b tr ca b E vat b kg a b D D D b b PA b rs b b b Fa a t a a a a t a a _ one b OE ae b Re ii b ate eee y be a ne ee eae b Attempting to compile such a program to chemical reactions using the standard DSD compiler will cause the compiler to loop indefinitely as it attempts to discover the entire reaction graph which is infinite It is recommended to use the JIT simulator described in the Simulation algorithms in Visual DSD section below when dealing with programs that involve polymer reactions To alert users to this problem the system produces an error message if a polymerization reaction is detected when the Polymers checkbox is unchecked as follows Compilation Error Parse error Species a lt t b gt and t b could interact to produce a polymer This error message also appears if an unproductive polymerization reaction is detected when the Unproductive checkbox is checked but the Polymers checkbox is not The States tab on the right hand side and its four sub tabs the Graph Text Summary and LBS tabs are populated with output when the States button is clicked This computes the continuous time Markov c
30. its run When the JIT simulator is running it checks after each reaction to see if the products of that reaction have been seen before if not a single compilation step occurs which augments the system with the new species and the new reactions which are made possible by the introduction of those species Thus the reaction graph and list of species is gradually built up over the course of a simulation run Running the same program multiple times in JIT mode may produce a different final reaction graph depending on which species were produced during each run Using the JIT simulator offers significant practical advantages when handling large scale systems such as those which include leak reactions In many cases simulating a system with leaks using the JIT simulator is comparable in speed to simulating that system without leaks using the stochastic simulator Furthermore the JIT simulator is essential when working with systems which have the potential to form DNA polymer molecules of unbounded size In these cases the JIT simulator only compiles the subset of reactions which are reachable from a species that has been created during the simulation run as opposed to the full infinite reaction graph Thus the JIT simulator may be the only way to run DNA polymer programs without causing the compiler to loop forever Populations and concentrations In Visual DSD quantities of DNA complexes are specified as molar concentrations which denote the num
31. mains UpperStrand lt Domains gt LowerStrand Domains In a double strand we assume that the domains listed between the square brackets are the domains on the upper strand so a double strand written as D1 D2 D3 would be visualized as follows where the small arrows at the top left and bottom right denote the 3 ends of the DNA strands Di D2 D3 Di D2 D3 DNA gates are a key aspect of the strand displacement computational mechanism A Gate is simply a concatenation of one or more Segments of DNA Gate Segment Segment Gate Segment Gate LowerStrand UpperStrand LowerStrand UpperStrand UpperStrand LowerStrand Overhangs Toehold LowerStrand UpperStrand Segment Double Double Overhangs Overhangs Double Overhangs Double Overhangs A Segment can take various forms 1 a single toehold domain on the lower strand t 2 an upper strand with a non empty list s of domains lt s gt 3 a lower strand with a non empty list s of domains s 5 4 adouble strand t1 t2 ti t 5 a double strand with overhanging upper and or lower strands on the amp left and or right 1b lt 1t gt s rb lt rt gt note that the order of the overhangs does not matter 15 By concatenating multiple segments we can produce DNA gates with complex overhanging structures The single colon represents concatenation along the lower strand while the double colon
32. nes Route Edges Options Y Fit Zoom 145 So m Concentration nM Time s In the Graph tab you can adjust the zoom level of the network display using the slider or the text box or use the Fit button to automatically select the zoom level which just fits in the entire graph The graph viewer has four available modes Pan where the mouse can be used to drag the graph around Zoom where you can drag out a rectangle to zoom in on Layout where nodes can be dragged around to modify the shape of the graph and Edge where edges can be added The Layout button on the right recomputes the original layout of the graph if it has been edited The dropdown labelled Options offers various choices for the layout Options gt Load dat Layered a _ Border Separation 5 4 _ Horizontal _ Aspect Ratio A _ Group Initial Nodes _ Show Rates Show Units Render Species The Horizontal checkbox toggles whether the graph should be arranged with reactions going from top to bottom or from left to right and the Aspect Ratio checkbox toggles whether the graph layout algorithm should attempt to match the aspect ratio of the window in some cases this can improve the readability of the graph The Export tab on the far right provides various representations of the species and reactions in the system For instance the SBML sub tab contains a d
33. nts 22 e The only characters allowed are upper case A C G and T e Only one sequence is permitted per line e The same sequence cannot be repeated e Toehold sequences cannot be longer than 9 nucleotides e Specificity sequences cannot be shorter than 10 nucleotides The tool does not perform any other sanity checking than this so further investigation of the sequences may be required to verify that they are appropriate for use in experiments If a large number of DNA sequences is supplied these sanity checks can slow down the compilation process The Check sequences option allows the user to disable the checks for repeated sequences and for the lengths of toehold and specificity sequences to speed up compilation These checks can be safely disabled for the preloaded sequences that come with the Visual DSD tool The sequences are output in order into the Domains tab when the Compile button is clicked If the program contains more domains than there are available sequences the system indicates this in the Domains tab as shown below The Domains tab also shows explicitly which end of the DNA sequence is the 3 end and which is the 5 end Compilation Simulation Analysis Species Reactions Graph Text Domains SBML Zoom i 3 gt 5 TATTCC 3 5 gt 5 GCTA 3 1 gt 5 CCCTTTACATTACATAACAA 3 2 gt 5 CCCAAAACAAAACAAAACAA 3 4 gt 5 CCCTTTTCTAAACTAAACAA 3
34. on stage can produce many thousands of reactions and take a long time to complete To mitigate this effect the system automatically filters out duplicate leak reactions i e those with the same reactants and the same products and also leak reactions which are a duplicate of a non leak reaction This helps to reduce the number of reactions in some circumstances without affecting the behaviour of the system too much but it is still possible to derived large numbers of leak reactions from a seemingly small system Note that Visual DSD does not currently support leak reactions which create DNA polymers by binding two gate molecules together Simulation algorithms in Visual DSD Simulation Method Stochastic Direct 7 Stochastic Direct Stochastic JIT Deterministic RK547M Deterministic Stiff BDF Deterministic RKF54 PDE Periodic id PDE Periodic 2d CME Integration Visual DSD provides a choice of different simulation algorithms The Stochastic Direct simulator uses the Gillespie algorithm to generate a possible trajectory of the system over time The Deterministic RKF54 simulator uses a non stiff ODE solver using a Runge Kutta Fehlberg method Fehlberg 1969 to produce a smooth deterministic plot of concentrations evolving over time In both the stochastic and deterministic simulations mass action kinetics are assumed such that the propensity of a reaction is proportional to the product of the reactant
35. ons Process Directives Process Process Declarations Process Directives and Declarations stand for possibly empty sequences of the Directive and Declaration non terminals which are described below Directives Directives are instructions to the Visual DSD simulator and data plotter Directive directive directive directive directive directive directive directive directive directive directive directive directive directive directive duration Float duration Float points Integer sample Float sample Float Integer scale Float concentration String time String plot Plots leak Float tau Float migrate Float lengths Integer Integer tolerance Float toeholds Float Float Duration directives tell the simulator how long to run for as a floating point number Optionally they may include a Points value with an integer value which specifies how many samples of the species populations to take during that time if none is provided the default is to sample species populations after every reaction which can produce a large number of data points The Sample 10 directives are an alternative syntax which has the same behaviour and is included for backwards compatibility Increasing the number of data points in the same period of simulation time produces more fine grained results but the display may be less responsive Similarly if the number of data po
36. require special attention to conventions The mass action kinetics for a T reaction k X A gt B can be understood in two distinct ways both appearing in the literature rx A k difference lies in whether one expects the factor k to be included inr or not Conventionally for Depending on how one interprets r the propensity of the reaction can be either r x A or the deterministic simulations one would expect the first formula while for stochastic simulations one would expect the second Thus the system will by default choose the formula based on the simulator choice as this is more likely to enable rate constants to be copied from books or papers However the choice can be forced via the directives kinetics deterministic and kinetics stochastic respectively State space analysis Visual DSD can also compute the graph of all reachable states for the system from the initial molecules This is useful for debugging individual components and for verifying larger completed designs Note however that if there are many possible interleavings of the reactions the state graph can be very large and this can take a long time to compute in such cases the computation can be stopped using the Pause button When the States button is clicked the system computes the continuous time Markov chain and populates the four sub tabs of the States tab on the right hand side Note that the state space analyser always uses JIT compilation as i
37. rocess Value constant Process Process Species new Name Value Value Process new Name Process Processes Processes Process Process Processes The first four lines of the grammar allow the populations of certain processes to be specified in each case the value should evaluate to an integer The constant keyword specifies that the population of a particular species should never change even if it participates in reactions which should consume that species We will discuss the grammar of species themselves in detail below The new process declares a domain which has local scope i e which can only be used within the body of that process As with globally declared domains these can have optional binding and unbinding rates attached otherwise default rates are used as described above This facility does not add expressive power to the language but makes it easier to organize larger programs by re using the names of domains without mutual interference Finally we can run multiple different kinds of processes in parallel using the vertical bar notation This is essential because molecules must be in parallel in order to react with each other If two sets of identical processes are placed in parallel the system will notice and add up their populations to produce a single population value for each species Species The Visual DSD language allows various kinds of DNA molecules to be expresse
38. s The deterministic ODEs are equivalent to the stochastic dynamics in the limit where the population counts approach infinity The difference between the plots can be seen by comparing the plots produced for the Catalytic example by the stochastic left and deterministic right simulators below Load data Seve data Resample Edit palette Show allitide all lt 2 3 4 gt lt 4 5 gt lt 12 2 lt 62 3 4 5 4444 4644 S OLEL AL 19 An interesting option is the Stochastic JIT just in time simulator Selecting this option affects not only the simulation algorithm but also the compilation of DNA species into chemical reactions The rationale behind the JIT simulator is that some systems can become very large with many thousands of possible reactions This means that the full compilation process can take a very long time This is particularly true for the larger example programs when leaks are enabled However since leak reactions have a low probability of actually happening we can spend a large amount of time computing reactions that will probably never happen during a particular simulation run The JIT simulator alleviates this problem by compiling new reactions dynamically during the simulation run When the CRN button is clicked in JIT mode only the Initial tab is populated with the initial species The other output tabs are populated when the simulation is paused or reaches the end of
39. s or floating point values The float of int and int of float functions explicitly convert between the two types thus offering a way round this restriction Alternative transition rules for Visual DSD Compilation Mode Default Infinite Default Finite Detailed The tool offers a choice between four semantic models which are selected using the Compilation menu These specify the reactions between DNA molecules which can take place Note that all binary reactions are mediated by complementary toeholds We summarise the choices as follows e Infinite molecules are identified up to branch migration Strand displacement toehold covering and toehold unbinding are assumed to have infinite rate Toehold binding is assumed to have a finite rate Note that unproductive reactions do not appear in the Infinite semantics because the toeholds will unbind immediately if no displacement is immediately possible This means that the Unproductive checkbox has no effect when the Infinite semantics is selected so it is greyed out when Infinite is selected This means that certain 17 programming idioms such as cooperative strand displacement are not possible using the Infinite semantics e Default molecules are identified up to branch migration Strand displacement and toehold covering are assumed to have infinite rate Toehold binding and unbinding are assumed to have a finite rate This semantics was described in Phillips
40. s which explicitly state the binding and unbinding rates of the domain respectively If these are omitted the default rates set by the toeholds directive are used if no such directive is present in the program the defaults are 3 0 10 nM ts t for binding and 0 1126 s t for unbinding It is worth pointing out that not all of the domains used in a program need to be declared globally in 12 this way If the system detects an undeclared domain in the program then it is assumed to have these default binding and unbinding rates This allows programs to remain short while allowing the flexibility to modify the interaction rates of certain domains as needed No units are supplied for the rate values it is up to the programmer to ensure that all rates are given in the same implicit units The unbinding rates are used in the Default semantic model see below Value definitions also written using def assign a value to a name Any subsequent uses of the name will refer to this value unless there is an intervening binding occurrence of the same name i e an instance of def or new The system will raise an error if a name is used without having first been bound to a value unless the name is used as a domain as mentioned above Processes The core of Visual DSD is a process calculus tailored to modeling DNA interactions The grammar of processes is as follows Value Process constant Process constant Value P
41. se a volume such that the number of individuals is a natural number The scale factor can be set by the scale directive where the default scale factor is 1 0 The choice of deterministic continuous or stochastic discrete simulation is also manifested in the units for the simulation plot The vertical axis of the plot has units of individuals for stochastic simulation and units of concentration for deterministic simulation Note that the units for rate constants are assumed to be consistent with the units for time and concentration For example if the unit for time are s and the unit for concentration are nM then the unit for the bimolecular rate constants are assumed to be nM ts t and the unit for the unimolecular rate constants are assumed to be s t Once a suitable scale factor has been selected in order to perform a stochastic simulation the molar concentrations are multiplied by the scale factor while the concentration dependent rates are divided by the scale factor For example if the scale factor is 100 nM then a 1 is converted to a stochastic rate of 0 004 s for concentration dependent rate of 0 4 nM ts simulation Additional details on converting between populations and concentrations can be found in Section 4 2 of Cardelli 2008 including specific conversion rules for homodimerization reactions Homodimerization is a special case of homopolymerization which is somewhat rare when dealing with DSD species but does
42. states the populations of species present in the initial state and any terminal states and the minimum and maximum populations of each species in the state graph The Summary tab presents a visual summary of the populations of species present in the initial state and any terminal states Under Export the PRISM tab displays code generated for the PRISM model checker which corresponds to the state space of the system This can be used to verify properties of the DNA circuit using probabilistic model checking Implementing domains as nucleotide sequences Visual DSD allows for domains to be compiled into nucleotide sequences providing a template for implementation of systems in DNA in the laboratory The DNA Sequences tab on the left hand side of the interface allows the user to enter the sequences that are used to implement toehold and specificity domains This tab is revealed by checking the Show Domain Editor box in the top level Options menu DNA Sequences DSD Code DSD Table CRN Code CRN Table CRN Graph Check sequences Reset TOEHOLD SEQUENCES TATTCC GCTA GTCA The windows for toehold and specificity sequences are preloaded with toehold domain sequences from the Appendix to Zhang Turberfield Yurke amp Winfree 2007 and specificity sequences which were generated by Qian for Qian amp Winfree 2009 The user can edit these and enter their own additional DNA sequences but there are some constrai
43. t If no Plot directive is supplied the system will sample the populations of all species at each time point Plots String Gate Strand sum Plots sub Plot Plot diff Plot Plot div Plot Plot Plots Plots Plots is a semicolon separated list of strand and gate species to plot If any of the species contain the underscore character _ this is interpreted as a wildcard which can match against a single domain causing multiple species to be plotted For example the pattern lt s t gt will match against lt u s t gt and lt v s t gt but not against lt u v s t gt Plot directives may also include quoted strings in this case the decision of whether to plot is based on an exact string matching on the names of species This can be useful when dealing with locally restricted domains as these are automatically renamed by the system It is also possible to plot simple arithmetic functions on species populations such as the sum of the populations of multiple species using the sum keyword followed by the species enclosed in brackets For a given pair of plot species P1 and P2 one can also plot the population of P1 minus the population of P2 using sub the difference using difft or the ratio of P1 to P2 using div The other directives allow the user to set the values of certain constants used by the system 11 e Leak directives set the rate of a le
44. t calculates the state space regardless of what is selected in the Simulation menu Simulation States Inference Verification Export Graph Text Summary LBS The Graph tab displays the set of reachable states as a graph using a similar graph display to that used for the chemical reaction network Each reachable state is represented as a node of the graph 21 and every possible reaction is represented as an edge There are some additional statistics along the bottom on the size of the state space and a checkbox to Draw images which toggles between graphical and textual views of the populations within each node The initial state of the system is highlighted with a thick black border on the left in the image below and any terminal states that is states from which no reactions are possible are highlighted with a thick red border on the right in the image below Simulation States Inference Verification Export Graph Text Summary LBS E a Pan _ Zoom _ Layout Layout Splines y Route Edges Options Fit Zoom 67 laj i I t Oo ah 0 0003 0 0003 Number of terminal states 1 Total number of states 9 Total number of transitions 15 The Text tab displays a textual summary of the state space with statistics on the size of the state space numbers of states transitions and terminal
45. t version and text DOT and SBML outputs are written to the filesystem 23 so that they can be opened in other tools It is also possible to run the simulator from the command line producing a CSV file which can be analysed in a spreadsheet application The command line version of the tool is available from the Visual DSD webpage in binary format for Windows and Mac OS X 10 5 as well as a platform independent Objective Caml bytecode file The Objective Caml programming language version 3 12 0 must be installed in order to run this version seehttp caml inria fr download en html for downloads and further information Binary distributions of Objective Caml are available for Windows Linux and Mac OS X and the source code distribution can be run on a wide variety of machine architectures using the Objective Caml runtime Once the correct version of the Objective Caml runtime has been successfully installed the command line version of Visual DSD can be run with the following command ocamlrun dna bytecode flags filel dna filen dna The dna files are a list of one or more files to be processed by the system The flags consist of zero or more of the following optional commands which mimic the behaviour of certain user interface elements from the Silverlight version of the tool e unproductive adding this flag includes unproductive reactions e leaks this turns on the leak model described above e polymers this enables polymer r
46. these are found in the DSD Table tab which will be populated after compilation or after clicking the button Table in the top row in the sub tab Directives Under the heading Simulation the dropdown labelled Method allows the user to select which simulation algorithm to use for producing plots see the Simulation algorithms in Visual DSD section below for more details Under the heading Compilation the dropdown labelled Mode allows the user to select between four different reaction models which range from simplistic and efficient to detailed and computationally expensive see the Alternative transition rules for Visual DSD section below for a brief explanation The rest of the elements under Compilation allow further configuration of the semantics of the Visual DSD language The Unproductive checkbox toggles the inclusion of reactions of repeated toehold binding and unbinding These reversible reactions go back and forth at a high rate and take up a lot of simulator time without affecting the overall result of the simulation greatly An example of such a reaction is the following not taken from the catalytic gate example y b a t t b L gt b This reaction is deemed unproductive because no other reactions can happen involving the molecule on the right hand side since none of the domains next to the toehold match up Removing these spurious reactions slightly reduc
47. ut may increase the computational cost of the compilation and simulation phases Leak reactions occur because the edges of double strands can sometimes fray up exposing the nucleotides at the end of the lower strand This allows a matching upper strand to bind onto the exposed nucleotides as if they were a toehold and displace the rest of the strand as normal For example in the reaction below the free upper strand and the bound upper strand are swapped Note that leak reactions are distinguished from normal reactions by a grey reaction arrow Y X Y z a y b y If the bound strand has a toehold at one end a leak reaction can still take place though the toehold must first flip up to allow the free upper strand to bind onto the frayed end of the normal strand as in the following examples 4 y t x y Z y a y t b YY nn si 9 4 t y x y Z y a t y b _ ra y t y 18 Note that the system disallows leak reactions where the bound upper strand has a toehold at both ends because in this situation both toeholds would need to flip up at the same time in order to allow the bound upper strand to leak We assume that the rate of these reactions is sufficiently slow that we can disregard them entirely Enabling leak reactions can cause the number of reactions in the system to increase considerably This is particularly apparent when the Finite or Detailed semantics are selected In these cases the compilati
48. y can bind to and unbind from their complementary strands quickly and easily The basic syntax of a toehold domain is simply a domain suffixed with a caret but there is also the possibility of explicitly complementing the toehold domain as described above Toehold Sequence Sequence The explicit complementation operator is a new addition to the Visual DSD language It was not included in the syntax of Phillips amp Cardelli 2009 because there it was assumed that the complemented sequences were all on the lower strand of the DNA molecule Our language provides for greater flexibility hence explicit complementation is required in the syntax To convert programs written in older versions of Visual DSD v0 12 or before to the new syntax simply replace every exposed toehold t by t We write Domains to stand for a non empty sequence of Domain or Toehold domain elements separated by a space Domain Toehold Domain Domains Toehold Domains Domains Domains are used to construct strands and more complicated DNA molecules An UpperStrand represents a single upper strand of DNA sequences and a LowerStrand represents a single lower strand of DNA sequences A Double represents an upper and a lower strand which are bound 14 due to Watson Crick complementarity as described above Since these are all just lists of Domains they must be distinguished by their ASCII syntax Double Do
Download Pdf Manuals
Related Search