You can also change the number in the "Max architectures" box. This defines a limit on the number of domain architectures that are shown on the graph. For some proteins the search can return several hundred, or even thousands, of nodes. The default setting is 150. It is only an approximate limit. The exact number shown depends on how many architectures lie within a cutoff "distance" that will give approximately this maximum number of architectures. To see all architectures, set this value very hign (eg 2000).

The radio buttons below the "Max architectures" box indicate whether all UniProt sequences are to be included or only those that are flagged in the UniProt database as "reviewed".

The button changes to "Plot Graph" which, when clicked, will restart the optimization process.

You can click-and-drag any of the nodes at any time.

Each node shows a set of coloured boxes representing the sequence of Pfam domains defining the particular architecture. Tall boxes correspond to Pfam-A domains, whereas small boxes correspond to Pfam-B domains. Some example nodes are shown below.

a.		b.		c.

The lines on the graph (ie the "edges" joining the nodes) show the relationships between the architectures, based on the similarity of their domain compositions.

The node corresponding to the architecture of the original search sequence is slightly larger than the others and has a grey background.

Navigation around the plot is as described on the left (taken from the Graph Criteria Panel (described later).

Of these operations, the most informative is left-clicking on a node. This provides information in the Data Panel (see below) describing the node's constituent domains and listing the protein sequences that have the given architecture. The data panel also identifies which sequences, if any, have whole or partial structures in the PDB.

Clicking the small black triangles between the panels allows you to close/open either panel. You can also adjust the size of a panel by left-click dragging the separator. Note that the Graph Criteria Panel cannot be reduced beyond a certain point.

The Print menu at the top right-hand corner of the frame gives you options for printing your graph.

The first option, Print graph, dumps the graph image to your printer. This may not always produce a particularly high-quality image.

The second option, PostScript file, writes a PostScript file. This is scalable and so can be used to generate high-quality prints or images, or even be converted to, say, a PDF file.

The first thing to do may be to click on the Freeze Graph button to freeze the graph's jiggling as it strives to optimize the relative placement of the nodes.

The Plot options radio buttons allow you to switch on either the UniProt sequences belonging to each Pfam architecture, or the PDB structures, or enzyme classifications.

These will then appear, attached to their parent nodes, as below:

a. UniProt sequences		b. PDB structures
c. Enzyme classes		d. Many satellite nodes

The nodes in example d are coloured red to indicate that there are too many sequences to display, so only a selected number are being shown.

Below the Plot options are two sliders. The top slider allows you adjust the typical lengths between the nodes on the graph to get a more compact or more distended plot.

The bottom slider is useful if you have a large graph with many architecture nodes. It can be used to prune some of the outer nodes.

Clicking the small black triangles separating the Graph Criteria Panel from the Graph Panel allows you to close/open either panel.

The Search tab at the top of this panel will replace it by the Search Criteria Panel, described next.

The large text box lists the all domains in the current sequence. It shows their Pfam identifiers, names and number of sequences. Initially, all are selected. Click on any domain to select it, and shift-click to add others to the selection.

The radio buttons above this list allow you to specify whether the architectures to be returned should contain all the selected domains (ie the AND option) or one or more of the selected domains (ie the OR option).

Note that, if you select all the domains and use the AND operator you will, of course, only get a single node on the plot: namely, that of the parent architecture.

The Organism button lists all the species for which proteins containing any of the above domains were found in the initial search. Each shows the species and number of protein sequences. Select which species to filter the sequences by.

The Sequences button allows you to limit the search to just those proteins for which there is a complete or partial 3D structure in the PDB.

Once you have made your selections, click on the Plot Graph button to retrieve the data and plot the graph.

Note that, if your selection criteria happen to exclude the parent sequence, it will still be returned by the search for reference purposes.