Exploring a model

A genome-scale model is too large to read as a list. The Explore section turns it into three searchable tables (metabolites, reactions, genes), and the inspector on the right shows any entity you click, with links to everything it is connected to. Walking from a reaction to one of its metabolites, to another reaction, and back is the main way to understand a part of the network.

The built-in view Active reactions after pFBA on iML1515: zero-flux reactions hidden, sorted by |flux|. 431 of 2,712 reactions carry flux in this solution.

1. Tables: Metabolites, Reactions, Genes, with their sizes (keys M, R, G).
2. Search (key F): ranked, across many fields at once.
3. Toolbar: Hide zero flux, sort, filters, columns, saved views, the subsystem browser and + Reaction / + Metabolite.
4. Subsystem browser: active / total reactions per subsystem; a click filters the table.
5. Table: virtualised, so it stays fast with thousands of rows. Click a row to inspect it, tick rows to act on them.

Search

The search box looks in several fields at once and ranks the matches:

Table	Fields searched
Reactions	ID, name, equation, participant names, genes, subsystem, EC numbers, annotations
Metabolites	ID, name, formula, compartment, annotations
Genes	ID, name, associated reactions

For example, phosphofructokinase finds PFK, PFK_2 and PFK_3 in iML1515. Ctrl+K opens the command palette, which searches the whole model at once and also runs commands such as Configure medium, Duplicate current state or Visualize active result on map.

Filter, sort and save views

• Hide zero flux (reactions), Hide zero turnover (metabolites) and Only genes with active reactions need an active result. "Zero" means below the zero threshold of the status bar. With only an FVA result, the reaction toggle becomes Hide blocked (FVA).
• Filters (reactions): |flux| above or below a value; reversible, irreversible or blocked; modified bounds only; reaction type (only or hide internal, transport, exchange, demand, sink); with or without gene association; favorites; compartment; subsystem. Active filters appear as removable chips under the toolbar.
• Sort: |flux|, flux, FVA width, ID, name, bounds or subsystem. Clicking the Flux column header twice sorts by |flux|.
• Columns: reaction, name, equation, LB, UB, flux, FVA min/max/range, subsystem, type, reversible, genes, EC, objective coefficient.
• Views store a search with its filters and sort. Five are built in (Active reactions, High-flux reactions with |flux| > 1, Exchanges, Transport, Modified bounds); Save current search & filters adds your own. They are kept per model.

Act on a selection

Tick rows (the header checkbox selects every row shown) to open the action bar:

Reactions	Metabolites
Add to Map · Block… · Reset bounds · FVA on selection · Filter to selection · Export CSV · Σ|flux| of the selection	Add to Map · Connected reactions · Export CSV

Right-clicking a row, a chip or a node of the local network opens the same actions for that entity: inspect, add to map, edit, set as objective, block, copy ID/name/equation/formula, create an exchange, sink or demand, add to favorites.

The inspector

Click any metabolite, reaction, gene or subsystem, in a table, an equation, a GPR rule, a search result or the local network, to open it in the inspector. Alt+Left / Alt+Right (or the arrows) move back and forward; the breadcrumbs show the path. Sections fold with a click on their title (Expand all / Collapse all at the top), and the ☆ star pins an entity to Favorites in the sidebar.

Reaction inspector for PFK after pFBA on iML1515.

A reaction shows:

• its equation, with metabolite names or IDs; every metabolite is a link;
• the flux in the active result, in mmol gDW⁻¹ h⁻¹, and the FVA interval and width if an FVA result is active (a note says when the values come from a result of another state);
• the bounds editor for the active state (see Medium, bounds and objective), with Block, Allow reversible and Reset original;
• objective coefficient, subsystem (a link), compartments and EC numbers;
• reactants and products with their coefficients;
• the gene–reaction rule, with each gene as a link and AND/OR highlighted;
• a local network (1 or 2 hops) and the raw annotations;
• actions: Add to map, Focus in map (when it is drawn), Set as objective, Copy, Edit reaction, Duplicate…, Delete reaction….

A metabolite shows its formula, charge, compartment, turnover, and the reactions that produce and consume it. This view is the core of metabolite analysis.

A gene lists its reactions, how many are active in the current result, and their rules. Gene knockout simulation is not implemented yet; the inspector suggests blocking the associated reactions to explore an essentiality hypothesis by hand.

A subsystem shows its reactions, metabolites and genes, with Add subsystem to map, Filter table, Select reactions and FVA on subsystem.

The local network

The small network at the bottom of the inspector is a navigation aid, not a pathway map. Around a metabolite, producing reactions are drawn on the left and consuming ones on the right: by the sign of the actual flux when a result is active, otherwise by the side of the equation. Around a reaction, substrates are on the left and products on the right. 2 hops adds the neighbours' other reactions; a column holds at most nine nodes. Currency metabolites (ATP, NAD(P)H, H₂O, H⁺, phosphate, CoA, …) are hidden unless cofactors is ticked. Click any node to move there.

Next: change the medium, bounds and objective, or go deeper into metabolite analysis.