MORO_USER MANUAL
Contents
1. First progress bar with a caption Appearance Ratio is used to limit the percentage of interactions of network which will be shown on the window Second progress bar with a caption Edge width is used to control the size of edge of network Final progress bar with a caption Node Size is also used to control the size of nodes of network Moreover we can choose the modules to be displayed by checking the boxes as follows R of ea ule amp amp ne Module s information 7 Module ID Nodes 4 l63 68 51 58 53 45 55 55 63 an ISISISISRIRIRICI lt 56a VT aAa aon es YES SV Ve Aele First the below figure is the result of the detail mode r a e e a as P gs a J ai gt Z A GE 42 ee a r E 4 j Fe ae Oo AS ep ee net ata a eae e Res CT PoS 3 A Ga x a A Br m A X Pe A SS E pat ae Poe kon lt 7 lt lt ge Te I Te b wd La p lt a X san k Yh rv AA A a x s AY 4 pS BSE APS NO SZ SG ARR ork ee i LN Y Sa T Na Pan Ws nA NaN p a T a Ore 7A e b KSA aie S lt o Finally the below is the result of the relative mode _Q Module Visualizat Detail Mode Brief Mode Absolute Relative Inhibiting type Activating type Appearance Ratio 0 50 100 Edge Width 0 50 100 0 50 100 2 2720 9f2695 1 1046 A W2B42 xj2. Y Source N X Interaction Y Target No IL1R1 O 1a BAX 14433 1 FKHRIDAF16G 17 AAG HSP90 O a ef 4EBP 1 1a Z 7I R_ 1 BARI 7T MR_ _ 14 INACTIVEG Lo Aa l a l For example The result of STKE network is shown as follows when we choose Organic in the Layout yFiles menu Ui ge AN z gt ie 3 The relationship between network dynamics and modularity Control panel Update Rules amp Initial State Settings Update Rule Scheme CONJ DISJ v Number of Initial States 10 Execute The correction between network dynamic and network modularity can be examined based on a set of initial state Here we provided an option below Number of Initial States 10 Note that The number of initial states should be larger than O value and smaller than or equal to all possible number of states 2 value n is equal to a number of nodes of network However it is very time consuming to analyse the network robustness with all possible number of states in large scale networks such as HSN or STKE Therefore we used 2n initial states to investigate the correction between network modularity and network robustness such as whole network in out module robustness Actually the outcome of 2 initial states is approximately similar to the result of 2n initial states because the numbers of found attractors are not much different to each other On the other hand if the network si
3. Modularity results The modularity values of HSN and STKE were computed to be 0 54778 and 0 73071 respectively On the other hand the mean s d results of the modularity in the Random HSN and Random STKE groups were 0 38095 0 06758 and 0 55145 0 05248 respectively The modularities of the real signaling networks were significantly larger than those of the random Boolean networks p value lt 10 b Robustness results The robustness values of HSN and STKE were computed to be 0 74967 and 0 672 respectively On the other hand the mean s d results of robustness in the Random HSN and Random STKE groups were 0 75282 0 00078 and 0 68028 0 00346 respectively The robustness values of the real signaling networks were significantly smaller than those of the random Boolean networks p value lt 10 _13 a network robustness 0 l l l l 0 0 1 02 03 04 05 06 07 08 09 network modularity random networks MHSN ASTKE N N 2 D 5 J 3 3 3 J 3 2 Z 0 8 4 t 4 fe 5 i oo gt p 0 7 1 mmm UUU K f 0 0 5 1 network modularity network modularity d 1 1 o o gt 209 wt eS 09 2 t sat 5 a 3 2 o AK o Q E 2 E 2 t 9 g 0 8 E08 5 0 7 0 0 5 1 0 0 5 1 network robustness network robustness Figure 2 Correlations between the modularity and robustness of 6400 random Boolean networks generated with V 50 an
4. RBI 82299070 VO MS anso A S wa 1 1800 M2320 i 3 410084 w a EN D60 SL aida Soon A Br Bac os a i 1928 4 4284 if PAPAR ROR Has 13654 T 4 Sd f SA nA REK AYS 1 3060 eS 3 aes 243S at BB X L ARU 2ds aaae A A hE NOI pza S oS ia ge aTe aan TRI Note that label of edge denotes the ratio of the number of interactions between a pair of modules to the maximal possible number of interactions between them that is w nn where w is the number of actual interactions between the pair of modules and n and n are the numbers of nodes included in each of the modules In out module robustness and centrality values of each module Results of each module amp amp whole network Module s information v Module ID Nodes Edges In Rob s In Rob r Out Rob s Out Rob r Degree Closeness Betweenness Stress Eigenvector 7 lo l63 85 0 3134 0 45037 0 4752 0 51564 18 0 02564 4 79365 20 l0 17086 A 7 ji 68 166 0 31248 0 4205 0 47885 0 51041 19 0 02439 3 30657 28 0 11851 J 2 51 62 0 31932 0 43418 0 47519 0 50913 12 0 02439 1 46667 8 0 14921 7 3 58 87 0 31731 0 39581 0 45756 0 49008 24 0 02857 30 16122 103 0 26761 E 7 j 53 79 10 31465 0 44134 0 45521 0 49449 20 10 02778 4 36429 34 0 29883 7 5 145 76 0 31334 0 50308 0 45278 0 47819 30 0 0303 119 79448 118 0 37575 7 l6 55 140 0 30511 0 43838 0 44839 0 48844 28 0 02941 127 23997 151 0 32987 7 7 55 86 0 3085 0 45998 0 45732 0
5. Network Analysis Batch mode Analysis par Setting Nodes E Help About Exit Plugin Choose Single Network Analysis This mode has a main control panel o For the relationship between network dynamics and modularity analysis Table Panel Update Rules amp Initial State Settings Update Rule Scheme CONJ DIS v Number of Initial States 10 Node Table Edge Table Network Table Analysis Parameters Analysis Results 2 Import network Imported networks can be a pre generated RBN signalling or regulatory network o Note that To import networks in this plugin the networks file must have at least 3 columns source node interaction type and target node where the interaction type only can be one of 1 1 and O that are corresponding to inhibition activation and neutral respectively For example a small part of STKE network is Shown as follows Source Node Interaction Type Target Node I IL1R1 1 IL1R1 14 3 3 1 BAD 14 3 3 1 BAX 14 3 3 1 FKHR DAF16 17 AAG 1 HSP90 4EBP 1 1 EIF4E 7TMR 1 B ARR 7TMR 1 INACTIVE G A20 1 IKK A20 1 TRAF2 A20 1 TRAF6 A ACTININ 1 ACTIN ABIN2 1 TPL2 ABL 1 CRK ABL 1 RAD9 AC 1 CAMP ACHE 1 APAF 1 ACHE 1 CYTC ACTIVIN 1 ACTRIIA B ACTRIIA B 1 ALK2 ACTRIIA B 1 ALK3 ACTRIIA B 1 ALK4 ACTRIIA B 1 ALK6 AIP 1 APAF 1 By using the built in import function of Cytoscape users can select a file in some formats and users have to specify which
6. and DISJ denote that each node of a random network would be assigned a conjunction and disjunction function respectively CONJ DISJ denotes that each node of a random network would be assigned a conjunction or disjunction function randomly o Choose Robustness against Initial state mutation if users want to calculate robustness against initial state perturbation value of each node of all random networks o Choose Robustness against Update rule mutation if users want to calculate robustness against update rule perturbation value of each node of all random networks 4 Module detection Step 3 Network s Modularity Module algorithm is applied to find all module of network Finally all results will be saved to text files 5 Batch mode case study In this section we present a case study for the new function of MORO batch mode simulation on RBNs A previous study Tran and Kwon 2013 showing that modularity and dynamic related robustness are negatively correlated significantly Here we use the batch mode function to validate the correlation between them First the network files of HSN and STKE are needed to be imported into Cytoscape software Next we execute two RBNs simulations by using the batch mode function To this end we choose shuffling techniques for the simulations Shuffle II for the first simulation o Shuffling intensity parameter is set to 4 For a simulation we set some following parameters Nu
7. structures of biological networks Bioinformatics 27 2767 2768 Leicht E A and Newman M E J 2008 Community Structure in Directed Networks Physical Review Letters 100 118703 Maslov S and Sneppen K 2002 Specificity and Stability in Topology of Protein Networks Science 296 910 913 Noack A 2009 Modularity clustering is force directed layout Physical Review E 79 026102 Shimbel A 1953 Structural parameters of communication networks Bulletin of Mathematical Biophysics 15 501 507 Tran T D and Kwon Y K 2013 The relationship between modularity and robustness in signalling networks Journal of The Royal Society Interface 10 Wuchty S and Stadler P F 2003 Centers of complex networks Journal of Theoretical Biology 223 45 53 _15
8. txt Usage of MORO will be demonstrated by following sections I Setting 1 Setup OpenCL options II Single Network Analysis 1 Load Single Network Analysis panels 2 Import a network 3 The relationship between network dynamics and modularity 4 Module visualization and displays of robustness and centrality value II Batch mode Analysis 1 Load Batch mode Analysis panels 2 Generate a network 3 Network dynamics analysis 4 Module detection 5 A Batch mode case study I Setting 1 Setup OpenCL options Choose Setting menu OpenCL Setting menu has three main options o CPU with one core only use one core of the CPU for tasks o OpenCL on CPU multi core use all cores of the CPU for parallelization of tasks o OpenCL on GPU device use GPU device for parallelization of tasks Methods CPU with one core DEVICE NAME GeForce GIX 680 C OPENCL on CPU multi core DEVICE VENDOR NVIDIA Corporation OpenCL on GPU device DEVICE VERSION OpenCL 1 1 CUDA DEVICE MAX COMPUTE UNITS 8 DEVICE MAX WORK ITEM DIMENSIONS 3 DEVICE MAX WORK ITEM SIZES 1024 1024 64 DEVICE MAX WORK GROUP SIZE 1024 NVIDIA CUDA GeForce GTX 680 AMD Accelerated Parallel Processing II Single Network Analysis 1 Load Single Network Analysis panels Run Cytoscape and MORO plugin will be automatically loaded in Plugins menu of Cytoscape as following Layout Apps Tools Help lex t App Manager meqaa s MORO gt Single
9. 49981 19 0 02941 8 10332 46 0 3666 4B 63 123 0 32145 0 43724 0 47255 0 51535 26 0 0303 10 72164 63 0 38044 _ Z lQ Lag 7 IN 3213272 IN_4A1604 IN 4629289 IN AQANA la NN IS in In IN 21102 j 2 This figure shows the detailed results of each module More specifically some values of each module are number of nodes number of edges in module robustness against initial state perturbation in module robustness against update rule perturbation out module robustness against initial state perturbation and out module robustness against update rule perturbation we also provide five centrality measures including degree Jeong et al 2001 closeness Wuchty and Stadler 2003 betweenness Freeman 1977 stress Shimbel 1953 and eigenvector Bonacich 1987 Finally the export button will export the result to a text file Whole network robustness Network s information Nodes Edges Robustness s Robustness r Modularity Modules In Rob s In Rob r Out Rob s Out Rob r 754 1624 0 672 0 66592 0 73071 16 l0 315 0 43456 0 46622 0 49938 3 This figure shows the summary result including the number of nodes the number of edges the robustness against initial state perturbation the robustness against update rule perturbation the modularity value the number of modules the in module robustness against initial state perturbation the in module robustness against update rule perturbation the out module robustness against ini
10. 9 Erd6s R nyi model Erd s and R nyi 1959 Erdos R nyi variant model Le and Kwon 2011 Shuffling model Maslov and Sneppen 2002 o Choose parameters to generate corresponding random Boolean networks o For the first three models users can set the probability of negative link s assignment this probability would affect the ratio of negative links to all links o For Shuffling model create random networks from a given network Choose Shuffle direction and sign of all interactions if users want to create random networks by shuffling the direction and the sign of every interaction from the given network Shuffle I Choose Preserve in degree and out degree of all nodes if users want to create random networks by rewiring the edges of the given network such that the in degree and the out degree of all nodes are conserved Shuffle II e Users can set the value of Shuffling intensity parameter The number of rewiring steps Shuffling intensity x number of edges o Choose Don t create view for random networks if users don t want to create view for random networks when the simulation is running If this option is selected MORO can save a significant amount of memory l 3 Network dynamics analysis Step 2 Network dynamics calculate robustness against initial state perturbation and update rule perturbation over a chosen set of initial states o Choose a update rule scheme for each node of all random networks CONJ
11. MORO A _ Cytoscape Plugin for Relationship Analysis between Modularity and Robustness in Large Scale Biological Networks Cong Doan Truong Tien Dzung Tran and Yung Keun Kwon School of Computer Engineering and Information Technology University of Ulsan 93 Daehak ro Namgu Ulsan 680 749 South Korea Department of Information Technology Center for Research and Development Hanoi University of Industry Hanoi VietNam User Manual As FOWO SOUP wrasitertaa aatreeap aang edue baeieae ine cuasas T E S 2 B HOVO U ee e E E 3 E o E 8 e E A ee N EE A E A E T ct E E E EA EE E E E o E AE E E T ee 4 l CEU OPERE OION era a E EES 4 IE Smrke Ne MOok ANYS Brerna a aa 4 tk Load Smell Network Analysis Panels ecenin EA 4 2 ADO OWO E oe a a a a a a a 5 3 The relationship between network dynamics and modularity cece ceeecceeeeeeeeeeeeeeeeeaeees 7 a Set update rules perturbations and states Of nodes cc cccceccccsecccneeeeesececeeseeaeeeeeeeseeseneees 7 b Whole network robustness examination cccccceecccceseeecceeeeeeceeeccecenscceseeeceeseeeceseeseeeaanees 8 Ca Nod l COTS CUO IN erases pete cause aueen anette eee eee ee aaa 8 d In Out module robustness examination cccccceeeccccseeccceseeecceeeeceeseeeceaaeeeeesesensaeseeenees 8 4 Module visualization and displays of robustness and centrality value cceeceeeeeeseeeees 8 amp Mod le VISU ANZ AION acc eses nocd ace
12. T Aapianchudenenisetenaceat 9 b In out module robustness and centrality values of each module ccecccseeeeeeeeees 10 Ca Whole nei Work TODUSUICSS oee r a 10 M Batch mode analys Bossa E a a E cebsueng 11 l Load Bater mode Analysis pael ecccsiinniei e a Ea 11 2 EE O NOE n a a a eo Be eee maconeccn at aa 11 Ds ANCIWOrK dynamics anal yS Seea ie hance me uchonlueaemiaahawaaaanieieenta 12 emmy oE Come E ol C6 0 a AE E een a ae ere Onin arte en T ee ee E tee ere ae eee et ae 12 2 Bate Me MmoOde Mase SLUCY canard densa e a E S A 12 A How to setup After executing the Cytoscape choose Apps parent menu gt Apps Manager Then a window will appear as a figure below Install Apps Currently Installed Check for Updates Download Site _http apps cytoscape org Search all apps 76 AgilentLiteratureSearch A apps by tag a View on App Store Install Next click the button Install from File to select the jar file MORO jar to install the MORO plugin OpenCL setup o Download Accelerated Parallel Processing APP SDK from AMD website http developer amd com recommended version is v2 8 and install it AMD Catalyst Install Manager Version 08 00 0907 AMD Catalyst Install Manager Version 08 00 0907 Analyzing System Welcome Installer Welcome Welcome Analyze Analyze Customize Customize Default Installation Locati
13. column is a source a target and an interaction type Edit View Select Layout Apps Tools Help Recent Session New gt Open Ctrl O Save Ctrl S Save As Ctrl Shift s fes Edges Import Network File Ctrl L Export Table URL Ctrl Shift L R Style Public Databases Alt L un Ontology and Annotation Print Current Network Ctrl P Quit Ctri Q gt Next in case of STKE txt file unselect the option space and check the option Transfer first line as column name start import row On the other hand in case of HSN txt file select the option space and check the option Transfer first line as column name start import row Select a Network Collection Network Collection Create new network collection v Node identiter Mapping Columa Ssharedname Interaction Definition Source Interaction Interaction Type Target Interaction z gt Column 2 4 lt gt Column 3 a A Columns in BLUE will be loaded as EDGE ATTRIBUTES Column 1 Advanced V Show Text File Import Options Text File Import Options Delimiter Preview Options Z Tab E Comma F semicolon F space E other Show all entries in the file Show first 100 entries Column Names Network Import Options V Transfer first line as column names Start Import Row Comment Line Default Interaction Preview Text File Left Click Enable Disable Column Right Click Edit Column
14. d 49 lt A lt 2031 a Relationship between network modularity and robustness the correlation coefficient was negative R 0 80303 p value lt 10 The results for HSN and STKE denoted by the rectangular and triangular points respectively were very close to the linear regression line b e Contributions of in module out module robustness to the relationship between network modularity and network robustness b Relationship of the network modularity to the in module robustness R 0 30383 p value lt 0 0001 c Relationship between network modularity and out module robustness not significant d Relationship of the network robustness to the in module robustness R 0 27801 p value lt 0 0001 e Relationship between network robustness and out module robustness not significant 14 References Barabasi A L and Albert R 1999 Emergence of Scaling in Random Networks Science 286 509 512 Bonacich P 1987 Power and Centrality A Family of Measures American Journal of Sociology 92 1170 1182 Erdos P and R nyi A 1959 On random graphs I Publicationes Mathematicae Debrecen 6 290 297 Freeman L 1977 A Set of Measures of Centrality Based on Betweenness Sociometry 40 35 41 Jeong H et al 2001 Lethality and centrality in protein networks Nature 411 41 42 Le D H and Kwon Y K 2011 NetDS a Cytoscape plugin to analyze the robustness of dynamics and feedforward feedback loop
15. mber of random networks HSN is 100 random networks and STKE is also 100 random networks Werecommend users to check the option Don t create view for random networks to save a significant amount of memory We also recommend users to setup OpenCL options for faster simulations by the parallelization of examining robustness Second BA model is used to generate 6 400 random networks with V 50 and 49 lt A lt 2 031 Result of simulation is shown in the following two figures Ji 60 4 Random STKE Random HSN 50 STKE HSN 0 73071 0 54778 m amp 4a fi Frequency W S 1 20 10 4 Yy O om DO D amp D Oo amp A 5 amp 6 OD SN w amp M amp o N Q Qa Ne Ni Nio or X o S Ne Ne Ne Na ae a RA AA D S Network modularity 80 tiRandom STKE Random HSN 70 HSN b 0 74967 na Oo STKE 0 672 40 Frequency 6 A a A OG OD AN DM BW WD A wD amp 0D a D gt oc OD AO 6 oO AP amp amp FF OB Qe BD NV WwW VY WY NY Ow OM DD OD OD WM SF FF LF FF SF amp S SS FM HFM FM FF NV WY WY A Network robustness Figure 1 a Modularity and b robustness of the real HSN and STKE signaling networks compared with those of random Boolean networks The Random HSN and Random STKE results were taken from 100 random Boolean networks that were generated by shuffling interactions of the HSN and STKE networks while preserving their degree distribution respectively a
16. on ste Install Install C Program Files ATI Technologi nsta Finished Finished http Avwav amd com Cancel http Avwew amd com ek f net gt f _ concet_ AMD Catalyst Install Manager Version 08 00 0907 Welcome Installation complete warnings occurred during Analyze installation View Log for details Customize Install Finished http www amd com o Install appropriate drivers for all Graphics cards of the computer Download example network files from http mo ro sourceforge net for a test Note Run the Cytoscape software with an Administrator permission for example in Windows 7 a user can right click on the icon of Cytoscape and choose Run as administrator because MORO needs a permission to write result files in some functions B Howto use We explain how to use by using a case study based on the following networks HSN for Single Network Analysis o The human signal transduction network with 5 443 nodes and 37 663 links after removing neutral interactions from the original dataset version 4 downloaded from www bri nrc ca wang o Use the file HSN txt STKE for Single Network Analysis o STKE consists of 754 proteins and 1 624 interactions after removing all neutral interactions from the integrated network by using the 51 canonical pathways not specific to any organism or cell type obtained from http stke sciencemag org o Use the file STKE
17. results about the whole network Detail Mode Module s information Brief Mode V Module ID Nodes Edges In Rob s In Rob r Out Rob s Out Rob r Degree Closeness Betweenness Stress Eigenvector Absolute l63 0 3134 0 45037 0 4752 lo 51564 18 0 02564 4 79365 0 17086 Relative 1 0 31248 0 4205 0 47885 0 51041 19 0 02439 3 30657 0 11851 0 31932 0 43418 0 47519 0 50913 12 0 02439 1 46667 0 14921 0 31731 0 39581 0 45756 0 49008 24 0 02857 30 16122 0 26761 iio AE i o a soe f 0 31334 0 50308 0 47819 0 0303 19 79448 0 37575 0 30511 0 43838 0 0 48844 lo 27 23997 0 32987 i ee oe ieee pean _ ie ponts Eun i aia jo 32145 0 43724 0 47255 0 51535 26 0 0303 10 72164 10 38044 IN_ 21272 IN A16NA IN AQANA i In N 21103 NNISS Se 98 V Inhibiting type V Activating type Appearance Ratio 0 50 100 Edge Width 0 50 Node Size Network s information Nodes Edges Robustness s Robustness r Modularity Modules In Rob s In Rob r Out Rob s Out Rob r 754 11624 0 672 _ 10 66592 10 73071 116 0 315 10 43456 0 46622 0 49938 3 a Module visualization There are three options for module visualization namely detail mode absolute mode and relative mode In the Inhibiting and Activating items we can activate them for displaying interactions of network If no options are selected network will not show any type of links
18. sis p27 pRb S phase genes X AlO Oll l O a l O O ad f ad M OO M MOOO b Whole network Robustness Examination After setting the update rule and the number of states we can examine the whole network robustness More specifically y G is the robustness against initial state perturbation and y G is the robustness against update rule perturbation The results will be shown on the table in panel In the next section we will show the results more detail c Module detection After the whole network robustness is examined the module detection algorithm Leicht and Newman 2008 is used to find all modules of a network and the modularity is calculated by using an optimization algorithm proposed in a previous study Noack 2009 In the next section we will show the result of all modules in network and three types of module visualization d In Out Module Robustness Examination After all modules are found in out module robustness Yin G and Yout G can be examined by using the Hamming Distance The detailed results will be shown in the next section Module visualization and displays of robustness and centrality values The below figure shows the main interface of module visualization and results displays of robustness and centrality values Part 1 is used to setup parameters before module visualization Part 2 shows the robustness and centrality values of each module Finally Part 3 shows the analysis
19. tial state perturbation and the out module robustness against update rule perturbation Besides this result will be saved to a text file by using the export button 10 III Batch mode Analysis 1 Load Batch mode Analysis panel Choose the menu Plugin MORO Batch mode Analysis This mode has only one main control panel Step 1 Random Boolean Network generation Step 2 Analysis of network s dynamical and topological pro Network dynamics Update rule Scheme CONJ DIS v ber of Nodes IVI Number of random States 1024 Number of Interactions added Number of random networks 1 at each step d Number of Initial Nodes e J Don t create view for random networks Save detailed results for each node of all networks Probability of negative link s assignment 0 5 Run simulation Wait for inputting parameters Analysis of Network dynamics and modularity can be performed by three following steps o Step 1 Random Boolean Network generation o Step 2 Network dynamics o Step 3 Network s modularity There is an option for saving results o Choose Save detailed results for each node of all networks if users want to save two types of robustness results of each node of all random networks Three next sections will describe details of the above steps 2 Generate a network Step 1 Random Boolean Network generation o Choose one of random Boolean network models Barabasi Albert model Barabasi and Albert 199
20. ze is not so large in general the number of nodes lt 20 we can examine all possible initial states a Set update rules perturbations and states of nodes Each node in the network has its own state and update rule Thus before executing the network dynamics analysis users have to set update rules and states of nodes e Right after a network is either generated or imported update rules and states of all nodes are randomly assigned But users can specify them manually e Regarding update rules we provide three updating rule schemes CONJ DISJ CONJ and DISJ o If CONJ DISJ is chosen users can either specify or randomize update rule for each node o Note that 1 and 0 denote disjunction and conjunction respectively e Users also can either specify or randomize state for each node o Note that 1 and 0 denote on or off state respectively In visualization e Nodes having state 1 and O are represented by gray and light gray colors respectively e Nodes having conjunction and disjunction update rules are represented by rectangle and circle shapes respectively The state and update rule of each node is also stored as node attributes and they are updated whenever they are changed either by setting values in Node Attribute Browser Tab manually or simulation of network dynamics The figure below shows an example of a small network CDRN Node Attribute Browser ID State Update rule cyclin D1 cyclin E Mito
Download Pdf Manuals
Related Search