Using the 05Aug30 model in PLA > cd > iop ... a blank PLA window appears. It has a thumbnail picture of the net in the upper right and a view of a piece of this net (whats inside the red border) on the left. On the lower right is a tabbed panel for navigation (find) by names of occurrences (located proteins and chemicals) and rules, and specification of queries (selections). There is also a menu bar and tool bar giving alternatives for navigation, specification and execution of queries. The dish menu for this model offers you 9 options. RRME, TwoWaysToActRac, ThreeWaysToActRacX, FASL-FASR, quiz are smallish and load fairly quickly AdrenerticR and TNF-TNFR are modest size models, may take a couple minutes to load. EGR-EGFR, Everything and everything are quite big. Be prepared to way several minutes for a picture to appear. For this demo, select ThreeWaysToActRacX from the Dishes menu. This model has three different rules that lead to activation (loading with GTP) of Rac1 Browse the net -------------- -- in the thumbnail drag the red square around click on a point of interest -- in the net click on a rule (rectangle) or occurrence (oval) full names and options for further action appears in the info panel -- in the Find tab type `rac' in the text box now and you will see the first matching occurrence name (Rac1-GDP) again and you see the next match. choose the second match (click on the Find button) and the view will center on Rac1-GTP (if you push the click button its like pushing Find and then clicking on the node). You can also scroll through the list of occurrences or rules and double click on an entry to center it in the view. -- to find the `other end' of an arrow, click on it. it turns red and offers you some options in the info panel, for example centering on the other end of the arrow. Doubling clicking on an arrow toggles which end is in view, when one is off screen. Find all the rules leading to Rac1-GTP (890,256,97) Subnets --------- You can focus attention on a subnet by specifying some goals and/or avoids using the Selections tab and clicking on the desired occurrences. Goals appear green and avoids appear orange in the view. Now clicking on the Subnet item in the horizontal bar will create a new window displaying the `relevant subnet'. This is the subnet obtained by removing all rules that have a connected avoid occurrence, and from the remaining rules selecting only those that can possible contribute to a path that achieves the goals. Once the subnet is displayed you can toggle between display as an independent net and display in the context of the original net using the `context' button in the horizontal bar. (Subnet elements are colored, rules and occurrences not in the subnet are white.) For example select Rac1-GTP goal EgfR-act avoid (or simply avoid Egf) (Selecting only the Rac1-GTP goal gives back essentially the full net! Avoiding Egf instead of EgfR-act gives a bigger subnet, but see below, the same path.) Find a path ------------- To make a path query you need to specify ONE or more goals and zero or more avoids. (If there are no goals specified you will get an empty answer.) If you have a subnet with at least one goal, you can use the Path(LoLA) or Path(Maude) buttons in the horizontal bar to find a path to the goal. (You can also send the query from the original net from which the subnet was obtained). There may be no path. If so, you should get an info window saying so. If a path is found, it will be displayed in a PLA window as the network of rules used in the path. Sometimes LoLA and Maude give the same path, often they give different paths, due to the fact that these tools use different strategies to search for a path, and return the first one found. Try finding a path in the subnet specified above? (Goal: Rac1-GTP, Avoid: EgfR-act) Which rule does it use? (256) You can also view the path in context. Now find another path. Keep Rac1-GTP as a goal, but allow EgfR-act and avoid Ia5Ib1-act. The subnet contains rules 97 and 890 but the path found is rule 890. Finally avoid Ia5Ib1-act and Vav2-act. Now rule 97 is used. This works because we have postulated Cdc42 is present in an activated state. There is a path using 97 with only Cdc42-GDP in the starting state, but it can't be specified in our current language. The problem is that in the given model, Vav2-act is needed to activate Cdc42, and if it is present, it is chosen to activate Rac1 avoiding the need to use rule 97. Future work includes developing a richer query language.