ProteoLens 1.0
Contents
1. A single external data source flat file or database table which is rich enough can give rise to multiple data associations Let us consider a hypothetical protein protein interaction PPI table that contains 3 columns protein IDI protein ID2 and confidence Each row of this table represents a putative interaction binding between two proteins and a confidence score i e derived from the number of independent observations for this interaction see Table 1 Such interaction table contains two conceptually different associations 1 interactions protein 1 gt protein 2 2 interaction scores interaction between proteins 1 2 lt entity An entity Bn Figure 2 Data association Table 1 A hypothetical interaction table score The ProteoLens data model Protein ID 1 Protein ID2 confidence enforces clear distinction between these 1 2 0 8 two concepts The first data association to 1 3 0 7 be created on top of Table 1 would 2 4 0 2 trivially treat each protein protein ID as a 4 5 0 4 separate entity and map first two columns 1 5 0 6 of the table as a data association The second data association must treat the protein Ids in the first two columns as one single entity interaction and the confidence score as the other entity i e 1 2 20 8 User interface requires specifying which column s of the raw data table constitute an entity and thus allows creating both da2. and can also make the application run out of memory Result menu Create data association Allows to create data association on top of the file File data preview window or database query Oracle data preview window In Oracle data preview window make sure that the query was executed before the invocation of Create data association some metadata from the executed query result is used by the application to build the data association Upon selection of this option selection dialog s will appear asking to further define the data association its name and the columns of the underlying table to be wrapped by the association Network View Popup menus Popup menus in a Network View are activated by right clicks Three types of popup menus can be invoked by right clicking on a node edge or in an empty area of the view node popup menu edge popup menu and view popup menu respectively Special type of node popup menu node selection popup menu is invoked when more than one node are selected and right click is performed anywhere in the view Node popup menu Show info A message dialog will appear showing the node s ID the number of adjacent edges in the currently displayed network not in the underlying network source and the values for this node of all the node annotations attached to the current view Delete node menu Deletes the node the popup was invoked on Once Deletes the node no further action taken Forever filter ou
3. gt graphics attribute values respectively This allows cutting out outliers in the data that stretch gradients over unnecessarily broad intervals and zooming into the regions of interesting changes in the data values Example log ratios changing from 10 to 10 but everything outside for instance 3 3 is differentially expressed what we really want to see are subtle relative expression levels of the genes within the 3 3 interval Then we would set the limits to 3 3 and everything outside will be shown with fixed colors Create new node annotation schema Available data associations Annotation display editor PPI ld1 lt gt Id2 z Protein ID lt gt GO annotation Treat data as Siung Protein ID lt gt avg score Filters Protein ID lt gt connectivity Protein ID lt gt gene Protein ID lt gt log ratio Applyfilterto original set current subset interaction lt gt score sot Search pattern Camy Ce Available node ID descriptors PrimaryiD Data values Data association has node IDs in noni Annotation display strategy For annotation visualization use Main label If multiple annotations use Use first x Display annotations as Oal accessible 8 displayed Set Reset ok 7 Press OK the current view will be re rendered Reference Menu Options Main Window File Menu Run Garbage Collection Use this option if memo
4. or shape annotation descriptor respectively Edges menu See Nodes Same functionality for edge default styles and annotations is grouped under this menu The only additional option is Show direction checkbox If checked all the edges are shown as directed with arrow on one side Direction of an edge is determined by the order of the pair of nodes in the network source Note that no consistency checks are performed multiple edges are currently merged and there is no distinction between directed and undirected edges Thus if multiple pairs describe the same edge in a network source or multiple network sources contain the same edge the edge direction will be determined by the first encounter of a node pair describing this edge Also if the option is turned on then all the edges will be shown as directed This option is provided as a simple temporary way to draw directed e g transcriptional regulation network alone combination of directed and undirected networks in one view will be supported in future version Filter menu Filter list menu Loading saving current filter list Load Opens Load File dialog List of filtered node IDs will be loaded from the selected plain text txt file the IDs in the file can be either comma or newline separated See Node Popup Menu Delete forever Expand with filter Save Saves the filter list from the current view into a newline separated plain text txt file Note filter li
5. seen as a visual query engine that displays only a part of the underlying network relevant to the particular problem under investigation Indeed in many practical situations the visualization of the whole interaction network for example tens of thousands of edges gives little insight if any and what is actually sought is a visualization of a particular pathway or interrelations among a group of differentially expressed genes etc When a Data Association is being attached to a view as network source it can be attached silently or with a concurrent query In the first case nothing will be displayed in the view as yet but the underlying network can be queried later In the second case only the entities from the underlying network that satisfy the specified query will be retrieved and immediately displayed in the view Note that the query is not persistent it is used to retrieve a subnetwork from the underlying network but the rest of the underlying network is not filtered out forever and can be queried later Only one network source can be attached to the view at any given moment Querying network source The most natural choice for the network source in our example is the original PPI network Let us attach the Protein ID 1 lt gt Protein ID 2 association built on top of Table 1 to a Network View and let us do it silently first Connect only option Note that the protein Ids from the Data Association are now interpreted as th
6. ProteoLens 1 0 User Manual Last Date Modified July 19 2005 Contents TCO DUC HON ri A EA a A AE E E E AEAEE ENARA 3 Concepts and USE CASES keien i E E E E a 4 SOL Wars ALEC nerina ar aea a RREA EEA ite aad 4 Data Associations AA A 5 Visualizing Data Associations s ccscsivenesisesdecesssecacosoticnseriaeancsses tessssenntsustecnlesssvaucdsended 6 IN SEDE SOUBCE cane A os 6 Querying HEMOS RA 7 Adding AMOO cos 8 Complex visualizations ies 8 Filtering and Attribute Querying dd 9 Reference Most Important Operations lt c c sccessseseccenedescossncdesesssstecsesetccsnacdernedetadentess 10 Create Data ASSOCIATION ii css 10 View and Delete Data Assocrationsy s csusesnecngindaun iia eae ae ass 11 Open Network e e oaair aan aas 11 Attach Network Source to the View this dialog will change s00M ooonooccincninncninncc 11 Create a Node Mantally 0 A A aan 13 Expand NOGE jie nba TAN EnS 13 Add Amin Otao ie sucess AS AR 13 Reference Ment Ops ida 15 Man WII it A 15 A ath ots e E E E E aa EEA AEEA 15 WANT LOAN aTe IAL EA EA E E A 15 Data Management MU nidad 15 NetWork VIEW ia 15 INCEWOTK ANSI statist unalone hos E E oun ty lem lag vacate acted bits aerate hod smash E a 15 O stata el Salsas ned Natl Plc ol ek te aula E e Sa acess 16 LA a E 16 Layout Menis did 17 ELO EL a 17 E AA O Eaa i 17 O ian a a a ech delete Ndi el aa ea nates alain fl oa eats 18 Data Preview Window flat files and database queries oooooccnin
7. Thus the query specified in the screenshot reads In the PPI idl lt gt id2 network find all the interactions where first interactor protein id 1 is one of the IN LIST 1 2 OR second interactor is one of the IN LIST 1 2 If OR were changed to AND we would be looking only for interactions among the proteins 1 2 including self loops Note also that such query does not return all the edges among all the returned proteins Suppose that interactions 1 3 2 4 1 2 and 3 4 are present in the network The query specified by the dialog shown in the screenshot will look up edges so that it will return 1 3 2 4 1 2 first interactor is 1 or 2 or second interactor is 1 or 2 Hence 4 nodes 1 2 3 4 will be imported into the network but the interaction 3 4 will not be imported If this is not the desired behavior and all the interactions among all the imported proteins must be shown check the Load interaction envelope checkbox or perform Update edges operation in the view after the network is loaded If Respect filter is checked the nodes which are in the current view filter will not be loaded into the view Set condition for subnetwork being loaded Available data associations Annotation display editor PPI Id 1 lt gt 1d2 Protein ID lt gt GO annotation Treat data as Seat d zut Protein ID lt gt avg score Filters Protein ID lt gt connectivity Protein ID lt gt gen
8. attached as annotations providing property values for nodes and edges Annotation descriptors associated with each annotation allow specifying visualization schemas values provided by an annotation can be displayed categorically using graphical attributes such as colors shapes line types etc and continuously for numerical value properties e g expression data using color gradients line widths and node sizes Concepts and use cases In this section the data model utilized by the ProteoLens application is described Multiple examples of using various types of data in different contexts are also provided The intent of this section is to concentrate on conceptual issues Thus we do not describe user interfaces here and do not explain e g how to create a data association step by step This information is provided in the Reference Every time such an operation is implicitly or explicitly mentioned in this section the appropriate section of this manual containing detailed instructions for the use of ProteoLens controls and dialogs is referenced in the left margin 4 Software architecture The software consists of two functional layers Data Association layer and Data Visualization layer Fig 1 The only common language spoken by these two layers is provided by Data Associations The purpose of the Data Association layer is to wrap external data flat files Oracle DB tables in a uniform way and present them as Data Associations Da
9. be displayed as fill color continuous range In addition we will 2 create edge annotation Protein ID1 Protein ID2 lt score The resulting visualization is presented in Figure 4 Few important things to note here are 1 the labels of the displayed nodes changed not their ids the ids of the nodes are still 1 2 3 5 2 multiple 2 redundant annotations are ignored MyGene3 is i specified as the name for Protein Id 3 two times in the table 2 only distinct annotation values count and the name is shown once 3 it is important to specify use all option if multiple annotations are expected in the case of MyGenel id 1 there are 3 different log ratios available from different probes and with the use all option they are shown all as a pie chart filling of the node 4 The annotations are permanently attached to the view until they are manually deleted if new nodes are brought into the view imported from a list created manually or brought in via a query All the currently defined annotations will be applied to them automatically and the new nodes will be rendered according to these annotations Finally annotations are independent from the network source annotations can be defined in a view which is not yet attached to any network in which case of course nodes can be only added manually or imported from the list there is no network to query Figure 3 Figure 4 Complex visualizati
10. ccnocococcconcnononconnccanonnnos 18 A nesses vesacse hin tata tents veedeastpaaciecs A E ANSE 18 QUEME 3 o a aC ee ae a 19 A uamert a emeam oat aitenaetaren oe seeeean a 19 Network View Pop p MENUS a TAS A A Messe eats 19 Node POPULAR 19 Node selection popup MEL A AS 21 Edge POPUP ment A 21 View popup MEM da 22 Filesystems View popup MEUS aid dada 22 Introduction ProteoLens is a data mining oriented tool for querying visualization and annotation of biological networks Its main features and capabilities include Flat files in tabular column delimited format can be attached as data sources Connections to Oracle databases can be opened and results of arbitrary SQL select queries can be attached as data sources Raw data can be previewed and input data can be narrowed to a relevant subset of columns from the original table Multiple Network View panes can be opened each View is a full fledged graph editor allowing to add remove move network nodes and edges and to change node and edge graphical attributes A data source representing a network can be attached to a view and visual querying methods can be used to retrieve and expand subnetworks from such network source an arbitrary data source can be interpreted as a network and multiple data sources can be successively attached to and queried from the same view which allows creating complex networks from multiple data sources An arbitrary number of arbitrary data sources can be
11. ckbox only for the filesystem filesystem tree nodes that represent files If checked the data in the file are assumed to be in variable width char separated format and when View operation is requested ProteoLens will attempt to read the data as table user will be asked to input the delimiter character Otherwise View will open a simple textual view of the file data associations can not be created from such views Unmount available only for filesystem root nodes Unmounts this filesystem root e g a database and all the descendant filesystem tree nodes Mount available only for the Filesystems root of the whole filesytems tree Opens Database Connection dialog and allows to mount an Oracle database as a filesystem under the root
12. d more nodes into your network through Load Data Association dialog otherwise your manual layout will be lost Organic layout orthogonal layout circular layout hierarchic layout radiobuttons Selected layout type will be calculated every time layout is requested by user or internally by the software the layouter engines are provided through the 3 party y files library Configure Opens Configure Layouter dialog Allows to fine tune the parameters of the currently active selected layouter Run Enforces re calculation of the layout in the current view using currently active selected layouter Find menu By node ID menu Finds and selects nodes with the specified IDs in the network in the current view IDs from the specified list that are not found in the network are silently discarded no special warning is given Specific IDs Opens an input text dialog where comma separated list of the IDs to search for in the network should be typed in IDs from file Opens Load File dialog IDs to search for will be loaded from the selected file flat txt format the list of IDs in the file must be either comma or newline separated i e a single long comma separated string or a single column with one ID per row respectively Store current selection The list of IDs of the nodes currently selected in the view will be internally stored can be later used by Expand operations The stored selections are not shared between diff
13. e Search pattern Protein ID lt gt log ratio Applyfilterto original set current su interaction lt gt score Camy Crese Data values GO_4 GO_2 GO_1 GO_3 Data association has node IDs in 8 Prote annot all e accessible OK Cancel 5 Conditional query A data association that maps Ids from the network being attached to some property i e GO group must be available Select such data association ProteinID lt gt Go annotation selected in the screenshot select which of the two columns of this data association maps onto the network entity Ids Data association has node Ids in The attribute values available for the entities in the network being attached will be shown in the Data Values pane select desired values GO_4 GO_2 selected e Click OK the Ids of the nodes that possess the specified attribute values GO4 GO_2 in the example above will be transferred to the text input box of the generic query see item 3 a generic query can be further formed using these ids as described above Create a Node Manually 1 The Edit gt Options gt Allow node creation box must be checked 2 Click in the network view a new node will be created 3 If this is a decoration node it can be left as is or tweaked into a figure caption for example 4 If this node represents a real entity an Id must be set right click on the node sel
14. e new association must be also specified here 8 If only two columns of the underlying table were selected the association is created each column would represent a separate data entity 9 Otherwise another dialog will pop up asking to specify which columns should be combined into the first or leftmost entity in the Data Association Key columns the remaining columns will be wrapped by the second or rightmost entity 10 Repeat steps 6 9 as necessary i e if more data associations have to be created from the same data source Data Associations are linked directly to the underlying data source The table views from which an association was created can be safely closed at any moment e View and Delete Data Associations In the main application menu use Data Management gt View Data Associations This opens a dialog that allows removing or renaming a data association If the association wraps a database query the associated SQL is shown and can be edited Update must be pressed upon the edition is finished Open Network View In the main application menu select Window gt New Window gt Network View Attach Network Source to the View this dialog will change soon 1 In the Network View window menu select Network gt Load gt from Data Association 2 In the dialog window that opens select network source in the upper pane PPI Id1 lt gt Id2 selected in the screenshot 3 If Con
15. e unique Ids of the not yet displayed network nodes Since nothing is displayed in the view we have to start the network querying process in some way In order to do that one or more entities should be created in the view One way to do this is to import a list of entities network nodes from a text file Each element of the list will be interpreted as a unique node ID multiple occurrences of the same ID will be ignored Note that it is the user s choice what particular ID to use for visualizations LocusID SwissProtID gene name local database primary key efc but the IDs should be consistent across all data sources i e if PPI network is specified in terms of LocusIDs then to be able to directly map expression data onto such network the user must have these data in the LocusID lt expression value format Another way to query underlying network is to create one or more nodes manually Note that manually created nodes do not have any ID assigned to them They are nodes without identities After creating a node an Id must be manually assigned to it through the node s pop up menu Let us create a node and assign to it Id 1 Having a node in the view we can query the network source for the node s neighbors Expand gt without filter option from the node s popup menu brings into the view all the 3 interaction partners of the node 1 which are 2 3 and 5 The result is shown in Figure 3 Note that if the node strictly sp
16. eaking node Id is not in the network then the expansion or any other querying operation will have no effect Nodes with Id 6 and Id MyGenel can be created but none of them can be expanded note that the YO Reference add annotation application does not know that MyGenel is the name of protein with id 1 MyGenel is not an id in the attached PPI network Adding annotations As it was already mentioned any Data Association can be used as annotation source Only the Ids of the nodes in the network are used for attaching annotations hence to make the annotation useful the association must map the Ids used in the network onto something else for instance the Probe ID lt gt log ratio annotation is of no use for us yet since probelDs are not present in the network First we will create friendlier names for our proteins In order to do that we will request a new Node Annotation For the annotation we will select Protein ID gt gene association and specify that this annotation will be used for the node labels As soon as the annotation is created the network view is updated and protein names are shown as the node labels Arbitrary number of annotations can be created as long as the visualization is intelligible We will create here two more node annotations ProteinID lt gt connectivity and request it to be displayed as an auxiliary label and ProteinID lt gt log ratio and request it to
17. ect Set Id from the popup menu 5 Note once set Id can not be changed switching one protein id for the other is considered to be a very bad practice Delete wrong nodes create new ones never attempt to change Id Expand Node 1 A view must be attached to a network source otherwise the operation has no effect 2 right click on the node select the expansion option from the popup menu see Reference Menu Options for the details Add Annotation 1 In the Network View menu select Visualization gt Nodes gt Add annotation or Visualization gt Edges gt Add annotation 2 A dialog appears screenshot for Node annotation shown 3 In the left top pane select association to be used for annotation 4 Using comboboxes in the bottom left select type of visualization to be used i e whether the values provided by the data association should be used for text label color shape etc 5 Note that multiple values can be associated with a single node in the network e If multiple annotations are allowed labels and pie chart node filling make sure that use all option is selected e Ifa visualization scheme can not support multiple annotations i e shape the effect of use all is currently undefined use use first instead though an arbitrary annotation value out of a few available will be displayed or use Max use Min for numerical properties 6 The annotations can be e asis such as lab
18. els the annotation values are used as text strings e categorical particular graphic attribute such as color or shape should be assigned to each annotation value of interest Many values can be mapped onto the same attribute value for example many different GO groups can be all required to be drawn as the same color or the same shape All annotation values do not have to be mapped onto visualization attributes empty attributes default are allowed e Continuos allows mapping numerical values such as expression values onto color size or width gradients Note that if Continuous annotation visualization is requested then the Data Values pane shows only lt lt MIN gt gt and lt lt MAX gt gt values for which the corresponding colors sizes or width should be specified the intermediate values will be drawn as gradients between these two values Important lt lt MAX gt gt and lt lt MIN gt gt specify the boundary values of the graphics attribute such as color between which the gradient will be drawn they are generally speaking not the data boundary values With the continuous visualization type Set Limits button becomes enabled that allows to specify where in the available data range the lt lt MIN gt gt and lt lt MAX gt gt colors or other attributes map to All the data values below and above the interval specified in the Set Limits will be collapsed onto lt lt MIN gt gt and lt lt MAX gt
19. en the Condition dialog that allows to select attributes i e GO terms of interest and only the entities from the network source that are annotated through the association D with the selected attribute value s will be imported into the view Reference Most Important Operations Create Data Association 1 In the filesystems pane left browse to the data source file or database table to enable database access right click on the tree root and mount the required databases 2 Ifthe data source if a file it must be in column delimited format any file can be pre viewed right click gt view but only delimited files can be imported and converted to data associations e Right click on the file check the Table data option Right click on the data source select View If the data source is a file specify delimiter symbol in the pop up dialog 5 The contents first few lines of the data source is displayed as a table e Ifthe data source is a database table then the SQL query is displayed in the top pane An arbitrary query can be entered here it should be executed first Query gt Run to retrieve correct metadata and the result of this query can be wrapped as Data Association see steps below 6 From this table pane menu select Result gt create data association 7 The dialog will pop up asking to specify which columns of the underlying table will be wrapped by the association the unique name for th
20. erent views and are not stored between sessions Clear stored selection Removes the previously stored list of node IDs note stored list is removed NOT the nodes from the network Visualization menu Nodes menu Node visualization and visual annotation options Default menu sets default node style All the new nodes manually created or automatically imported into the view first receive default display style Set Opens Node Styles dialog All newly added nodes will receive specified default style nodes already present in the network will not be affected Set Update Opens Node Styles dialog All newly added nodes will receive specified default style and default style of nodes already present in the network will be updated Add annotation Opens Create New Node Annotation Schema dialog Through this dialog an arbitrary data association can be attached as the node annotation source and annotation descriptor specifying how to display graphically this particular annotation can be created Edit annotation Shows the list of all annotations currently attached to the nodes in the network and allows removing annotation changing annotation descriptor parameters note it is impossible to change the type of visualization through this option for example to switch from using shapes to using colors for a particular annotation but for instance used colors or shapes and their assignments to particular values can be changed for a color
21. he node ID with the value currently displayed as the node label Use of this option is not recommended as it can easily lead to breaking the network integrity Expand menu Finds the neighbors of this node in the currently attached network source and brings them into the currently displayed network Note network source must be attached otherwise this operation has no effect With filter The filter will be applied out of all the neighbors of this node only those will be returned that were not previously filtered out see Filter gt Filter list gt Load in Network View menu section and Delete node gt Forever filter out in this section Without filter All the neighbors of this node in the currently attached network source will be retrieved To All with filter To stored selection with filter This option changes depending on whether the view has a previously stored selection If it does not then the first option is active and the group see below is defined as all the nodes of the current network If selection was stored see Find Store current selection in Network View menu section then this menu option automatically switches to its second version and group is the stored selection When this menu option is invoked the neighbors of this node which are also neighbors of any node in the group are selected and those that are not filtered out are imported into the view This option is thus finding all the paths of length 2 fro
22. ing data into the view From file Opens Load File dialog Through this dialog previously stored network gml format or list of object IDs i e gene or protein IDs can be loaded into the current view The current contents of the view will not be erased new data will be appended to the existing network From data association Opens Load Data Association dialog This dialog allows selecting a data association to attach to the view as the current network source and automatic extracting of a subnetwork from a data association into the view based on selection rules specified through the dialog controls Save as Opens Save File dialog The dialog allows saving either the complete network in the current view as gml file or list of all object node IDs found in the network in the current view as a one column txt flat file Save selection as Opens Save File dialog Only the current selection in the view will be saved into a file and only the option of saving node IDs from the selection into a one column txt flat file is currently supported Save image as Opens Save File dialog Complete display of the network in the current view regardless of the zoom level and display window boundaries will be saved in either JPG or PNG graphics format It is recommended to set zoom level to 1 before saving the image Print Opens Print dialog Through this dialog complete network in the current view regardless of the zoom leve
23. l and display window boundaries can be sent to a printer The picture will be automatically downscaled if needed to fit the page Close Closes the Network View Edit menu Options menu Sets graph editing options for this view Allow node creation checkbox If checked clicking on the empty area in the view creates a new graph node Otherwise clicks in the empty area have no effect Allow edge creation checkbox If checked clicking on the node in the view and immediate dragging the cursor creates a new graph edge to select the node mouse should be clicked and released in this editing mode If unchecked click and drag on the node selects this node and moves it after the cursor Zoom menu Zoom in Zooms the view display area into the network Zoom out Zooms the view display area out of the network Zoom factor Allows setting the zoom factor manually Zoom factor of 1 is fixed by a display size of the default node as compared to e g fitting the whole network into the display area so that the size of the network in the current view is not taken into account Fit network Set zoom factor automatically to make the whole network in the current view visible in the display area Layout menu Auto layout If checked loading a subnetwork into the current view from data association will automatically start layout calculations for the whole network in the view Disable this option if you have manually edited the layout and intend to ad
24. m a given node to the group not passing through filtered nodes Note all paths of lengths 1 or direct edges can be retrieved by Update edges operation see below To All with filter To stored selection without filter The same as To All with filter To stored selection with filter but the filter is ignored all the length 2 paths from the node to the group will be retrieved and added to the network in the current view Node selection popup menu This menu is invoked when more than one node is selected in the view and right click is performed anywhere in the view All the operations invoked through this menu apply to all nodes in the selection Delete Deletes all the selected nodes from the view Note that there is no Delete forever operation here as it is considered to be important enough to deserve careful one node at a time execution Set node attributes Opens Node Attributes dialog the graphical appearance attributes set through this dialog will be applied to all nodes in the selection Label ID Executes LabelsID for each node in the selection see Label gt 1D in Node popup menu section Expand selection menu Allows to retrieve and import into the current view nodes which neighbors of any node in the current selection Network source must be attached to the view or this operation has no effect Note that in typical networks full expansion of a group of nodes often results in a very large and unmanageable number of
25. nect only is checked the selected association will be attached as a network source but no querying will be performed at this time and nothing will be imported into the view 4 Generic query Filter data e If there are few known protein Ids to query for or to exclude type them into the text input areas comma separated 1 and 2 are typed in in the screenshot e If these are the ids of the entities to search for leave the default IN LIST values of the comboboxes e If these are the ids of the entities to be excluded set combobox es to NOT IN LIST e Set the concatenation combobox to OR or END e By default the two text areas are synchronized checkbox which means that only one should be filled the typed data is automatically mirrored into the second text area Select network source PPI 1d1 lt gt Id2 Protein ID lt gt GO annotation Protein ID lt gt avg score Protein ID lt gt connectivity Protein ID lt gt gene Protein ID lt gt log ratio interaction lt gt score Filter data protein ID 1 IN LIST vi 2 OR v y Synchronize lists protein ID 2 IN LIST Y 111 2 Condition _ Load interaction envelope v Respect filter _ Connect only OK Cancel Note that the query searches for the edges with the desired properties not for separate proteins present in the network
26. neighbors so there are currently no simple Expand operations implemented for selections in ProteoLens To All with filter To stored selection with filter same as Expand To All with filter To stored selection with filter in Node popup menu section except it selects all the nodes which are simultaneously neighbors of any node in the selection and of any node in the group and which are not filtered out all length 2 paths from selection to the group not passing through filtered out nodes To All with filter To stored selection without filter The same as To All with filter To stored selection with filter but the filter is ignored all length 2 paths from selection to the sroup Edge popup menu Show info A message dialog will appear showing the IDs of the nodes connected by this edge this ID pair is also the edge s ID and the values for this edge of all the edge annotations attached to the current view Delete edge Deletes the edge from the view Set attributes Opens Edge Attributes dialog that allows editing appearance of the edge line color width and line style can be manually set from this dialog Note automatic annotations attached to the view through Create Edge Annotation Schema dialog override manually set graphical attributes Save skin Saves the skin graphical attributes of the edge See Node popup menu Save skin Apply skin Applies previously saved skin to this edge If disabled then n
27. ng the probes onto the genes Filtering and Attribute Querying Other features to be mentioned here are filtering and attribute querying Filtering is currently supported by the node Id only Namely if some node is not desirable in the view a false positive a protein unrelated to the pathway under study etc then it can be deleted with the option forever In this case the node Id will be stored in the filter list and Expand gt with filter operation will never return such node the expand gt without filter operation overrides the filter and returns all the neighbors of the node in the currently attached network source The filter list is local for the current Network View i e it is not shared between the views However the filter list from a view can be saved into a file and then imported into another view or loaded when a new session is started Attribute querying is supported through the Load Network gt from Data Association option Although this option is used to attach network source it is safe to re attach the same Data Association as the network source again and again so that this option can be also used for querying If a data association D with an annotation i e Protein ID gt GO is defined in the current session it can be used to query the network being attached Namely instead of using the native query interface which allows only specifying Ids of the nodes to be imported the user can op
28. ntains the average confidence score for all the interactions involving given protein Note that if Table 1 was stored in the database the numbers presented in Table 3 could be calculated on the fly by application of an appropriate SQL query and the appropriate data associations could be created within ProteoLens on top of Table 1 If data of Table 1 are stored in a flat file then Table 3 must be pre calculated externally summary statistics calculation is not currently supported in ProteoLens Two associations created from Table 3 are Protein ID lt gt connectivity and Table 4 Differential expression Protein ID gt average interaction score measurements The last table we consider contains expression i data for our hypothetical proteins 1 5 The data Probe ID Protein ID Log ratio presented in Table 4 contain a hypothetical AA 1 0 8 microarray probe IDs a mapping of these probe BB222 2 0 6 ids onto gene ids and measured log ratios A CC333 1 1 4 Protein ID lt gt probe ID Protein ID lt log DD444 3 0 9 ratio and Probe ID lt log ratio associations EE555 4 0 1 can be created from these table Not all these FF666 1 1 3 associations might be necessary depending on GG777 5 0 05 the particular problem and visualization scheme a researcher has in mind For now we assume however that all 3 associations are created Visualizing Data Associations To proceed wi
29. o skin was previously saved View popup menu Show info Invokes a message dialog which shows the number of nodes and edges in the network Update edges Retrieves from currently attached network source and imports into the view all the edges relations among the nodes currently present in the view If there is no attached network source the operation has no effect Apply filter Removes from the current view all nodes found on the filter list i e Expand without filter was performed unadvertedly instead of Expand with filter or a new filter was just loaded from a file and the network in the current view has to be filtered against it Filesystems View popup menus Expand Expands the tree of child objects defined in a filesystem subdirectories files in a filesystem filesystem or schemas and tables in a database filesystem Same as left clicking on the tree handle on the left of the icon Collapse Collapses the tree of child objects Same as left clicking on the tree handle Refresh Refreshes the tree below this filesystem object reread the list of subdirectories and files in a directory of a filesystem filesystem or refresh the list of tables in a database file system View only for the filesystem tree nodes that are not sub directories but can contain data i e files in a filesystem filesystem and tables in a database filesystem Opens Data Preview window populated with the data from this filesystem object Tabular data che
30. ons Only one network source at a time can be attached to a view However new source can be attached to the existing view at any given time previous source will be automatically detached All the nodes and edges imported from the old source will remain in the view which allows for creating complex visualizations It is possible to combine in this way few different sources of interaction data or add transcriptional regulation or coexpression relevance network data from another source efc the current limitation is that multiple parallel edges are not supported hence in the situations when the overlap is expected it is recommended to use edge annotations instead Here we will demonstrate however another use of this capability Suppose that we want to add GO annotations from Table 2 and maybe few other annotations We still have the node shapes available for using for automatic annotations exercise Figure 5 attach GO annotations and display them as node shapes categorical but the nodes become visually overloaded In some cases it may be preferable to display annotations especially such as belonging to some group as a graph as well Let us keep the current annotations expression levels as the fill colors and attach Protein Id lt gt GO Data Association as a network source rather than an annotation Set default node shape to rectangle and expand all the protein nodes since the currently attached network sou
31. rce associates Protein Ids with GO terms the new nodes brought into the view by such querying will be GO terms and the Ids of these nodes will be the GO_1 GO_2 etc as provided by the Data Association In addition to displaying the fact of belonging to the GO group in a graph based way this representation also allows for yet another way to query for particular terms if we now expand the GO_2 node in the network we will obtain all the proteins associated with this term Then we can reconnect to the PPI network and query it for all the interactions edges among the currently displayed nodes The result is presented in Figure 5 Note that nodes for which an annotation is not available e g Protein ID gt log ratio annotation value is not available for the node with ID GO_ 1 are rendered with default graphical attributes With the associations previously specified it is also possible to decouple expression data from the protein nodes by using Probe ID lt gt Protein ID association as a network source drawing probe nodes in a graph based way connected to the corresponding protein nodes and using ProbeID lt log ratio association to color probes according to the expression values rather than proteins Such visualization might be helpful especially if a few probe level parameters are available and more detailed look at the this level is required still maintaining the higher level information about mapi
32. ry problems are suspected Memory status is displayed in the status bar in the lower portion of the main window This command directly requests Java Virtual Machine to perform garbage collection Save session Opens the Save Session dialog All the data associations currently defined in the ProteoLens session will be stored to the XML file selected through the dialog Load Session Opens the Load Session dialog Data associations previously saved in the XML file selected through the dialog will be recreated and imported into the current ProteoLens session Data associations are identified by their names if the name of data association being read from a file already exists in current session the error message will be displayed and this data association will not be read in If a data association is defined on top of the Oracle database SQL query and the required database connection is not yet open in the ProteoLens session the Database Connection Dialog will open Exit Exits the application Window menu New window menu Opens new window in the application desktop Network view Opens new Network View Newly opened view is not attached to any data source Data Management menu View Data Associations Opens the Data Association Viewer dialog that lists data associations available in the current ProteoLens session shows their properties and allows renaming deleting data associations Network View Network menu Load menu Allows load
33. sts are not shared between different views and are not saved by default between the sessions They must be loaded saved explicitly in each view Data Preview Window flat files and database queries File menu Load SQL Oracle data preview window only Opens Load File dialog Through this dialog SQL query previously saved as a standalone text file default is sql can be imported into the query pane of the window note this allows loading queries created in different applications such as SQL Navigator Save SQL Oracle data preview window only Opens Save File dialog Allows saving the query from the query pane into a standalone text file Close Closes Data Preview window Query menu Execute Oracle data preview window only Executes the query from the query pane of the window in the database and schema the window is attached to If query execution failed then the exception message will popup with short explanation of the problem as provided by the Oracle database JDBC driver see Oracle documentation for error codes If the query executes successfully first 10 rows of the result will be retrieved to populate the data preview pane in the window Execute amp Retrieve all Oracle data preview window only Use with caution This is the same as Execute but all the query results will be retrieved to populate the data preview pane If the returned table is very large it can result in unnecessary network connection overload
34. t Deletes the node and adds its ID to the filter list associated with this view The node filtered out forever will not be allowed to reappear in the view as the result of Expand operations performed on other nodes or when a subnetwork is loaded into the view by Load gt From data association operation Set attributes Opens Node Attributes dialog that allows editing appearance of the node fill color shape size label text label font outline color width and line style can be manually set from this dialog Note automatic annotations attached to the view through Create Node Annotation Schema dialog override manually set graphical attributes Set node ID This option is enabled only if the node does not have an ID for instance it was manually created Once the node receives its ID the option is disabled it is not allowed to change node ID If enabled this option invokes a text input dialog where the ID to be assigned to the node should be typed in Save skin Saves the skin most of the graphical attributes of the node Skin can be viewed as a rubber stamp that can be later applied to other nodes to make them look just like this one Note saved skins are not shared between network views Apply skin Applies previously saved skin to this node If disabled then no skin was previously saved Note saved skins are not shared between network views Switch layer Experimental no effect Label ID enforces replacement of t
35. ta Association layer user interface allows user to specify which particular subsets of data from which particular sources should be available to the application Data Visualization layer accepts only Data Associations and allows both drawing them as networks graphs or visualizing them as graphical attributes of the network nodes and edges Any available Data Association can be used in either of the two ways The application makes no domain specific assumptions about the nature and meaning of the provided data which leaves the user with responsibility of using right data at the right place but also allows for very high flexibility Data Data Association Visualization Database layer layer Data draws Data Associations El gt associations IM gt Local filesystem Figure 1 Software architecture overview YO Reference create Data Association YO Reference create Data Association Data Associations Data Association is the fundamental concept used in this software tool It provides the interface between external data and visualization Only Data Associations are accessible from visualization layer B A data association is a two way many to many relationship between data objects referred hereafter as entities Entities in turn are context and problem specific and unbreakable Hence a data association can be thought of as a table with exactly two columns Values in each column are entities Fig 2
36. ta associations described above see Reference To make the situation more interesting let us assume that few additional annotations are available First let the gene names and GO annotations be provided in a separate file Table 2 A hypothetical annotation table Protein ID gene annotation i 1 MyGenel GO 1 Table 2 Note that the table 2 is not 7 MyGene2 GO 2 normalized i e it contains redundant gene 3 MyGene3 GO 1 name annotation for protein ID 3 MyGene3 3 MyGene3 GO 3 in both rows The two obvious associations 4 MvGene4 GO 2 that we create from this table are Protein 5 MyGeneS Go 4 IDegene and Protein I D lt annotation Both associations make use of only two YO Reference create Data Association YO Reference create Data Association YO Reference open Network View columns of the original table and trivially Table 3 A summary statistics table wrap contents of each cell as an entity Note Protein ID connectivity Avg score multiple annotations for protein 3 and 1 3 0 7 multiple proteins 1 and 3 2 and 4 2 2 0 5 annotated with the same term 3 1 0 7 The next table that we consider here Table 4 2 0 3 3 contains summary statistics for the 5 2 0 5 interaction table shown in Table 1 The connectivity column contains the number of interaction partners for each protein in the network and the avg score column co
37. th our example let us assume that all the data associations described in the previous section are created and thus the visualization layer sees them all note that it never sees any raw data The main component of the visualization layer is a Network View Network View is a fully functional graph editor manual drawing and rearranging the nodes and edges in the network editing graphical attributes such as colors shapes etc that is capable of displaying automatically the data provided through Data Associations Network source An arbitrary data association can be attached to the view as a network source In this case the pairs of entities rows see Figure 1 provided by the Data Association are represented as pairs of network nodes linked by an edge The string literal contents of the entities underlying single cells of the data table or multiple cells concatenated YO Reference attach network source to the view Y Reference create a node manually Y Reference expand node 66 99 wit if an entity wraps a few columns of the underlying data will be taken as unique node ID The unique ID of every edge in the network will be automatically defined as id1 id2 where id1 and id2 are ids of the nodes linked by this edge It is important that a view does not have to show the whole attached network and that in most practical situations 1t will not be asked to although it can do this efficiently Instead a view must be
Download Pdf Manuals
Related Search