The scientific question you want to address should be clear, as should the message you want to communicate to the reader. Specifying your message will help you choose what to omit, what to represent, and how to represent it. For example, when representing a network, you should think about the type of pathway you want to model: Do you want to say that phosphoinositide 3-Kinase (PI3K) activates protein kinase B (AKT, signaling information), or do you want to say that PI3K transforms phosphatidylinositol 4,5-bisphosphate (PIP2) into phosphatidylinositol (3,4,5)-trisphosphate (PIP3) and that PIP3 binds to AKT (metabolic information)? The understanding of the network’s biochemical and biophysical organisation will help you structure the map, and SBGN helps you to sketch it out.

Keep your target audience in mind: different readers perceive different messages and focus on different aspects of a network. For example, a biochemist will generally be interested in a detailed map highlighting biological entities (e.g., macromolecules depicted in an SBGN PD map) and their interactions; whereas a cancer researcher will be concerned about mechanisms and feedback loops that occur in a specific cell type (depicted in SBGN AF notation). To ensure the successful transmission of your message, identify your audience and adapt the network's representation accordingly. Ask yourself: What do they know and what do they not know? What are they interested in? This will lead you to determine the most appropriate level of representation to visualise your network.

Design your map in a reasonable level of detail. Fig 1 shows one biological system represented with glyphs from the three SBGN languages. In SBGN PD, we represent the fact that the depolarisation of the membrane triggers the opening of the sodium ion (Na+) channel, which enables the import of Na+ from the extracellular space to the cytoplasm. In SBGN ER, the depolarisation stimulates the assignment of value “true” to the state “open” of the Na+ channel, necessary to stimulate the relocation of Na+ into the cytoplasm. In SBGN AF, a cascade of signals is shown where the depolarisation stimulates the activity of Na+ channels, which in turn, is necessary to stimulate the activity of Na+ in the cytoplasm. For more details about the symbols used, please refer to the SBGN reference cards (http://sbgn.github.io/sbgn/templates). Deciding what knowledge to transport through your map will guide you in choosing the right SBGN flavour.

List the reactions constituting your network and carefully choose the names of the biological components that will be displayed as labels in the SBGN nodes. The suitable SBGN glyph to represent a biological component can be chosen by using either the reference cards (http://sbgn.github.io/sbgn/templates), the information provided by your SBGN-compliant software, or the information given in the language specifications. For example, if your component is a protein, you may use the SBGN macromolecule glyph depicted on the reference card for SBGN PD. Once your biological components are mapped on SBGN glyphs, you should choose appropriate arcs to link the components together and build the connectivity of your network.

A network can be visualised at different levels of detail, as exemplified in Fig 2.

The SBGN does not tell you how to represent a system, but tells your readers how to interpret what you have drawn. It is important to be as specific as you can be about the encoded biology without diluting your message. In particular, omitted information might make it hard for others to interpret your network and hence, to understand your message. At the same time, the right level of abstraction must be chosen for the important parts of your network to stand out to the readers.

A careful design of your SBGN map could provide additional biological information on top of your network. Start by creating the components and their interactions (as mentioned in Tip 4), then add necessary information in respect to Tip 5.

A good layout improves the readability of your network and can also speed up interpretation, in particular if you followed Tip 2. For instance, people working on signaling pathways generally put the plasma membrane on the top of a map and the nucleus on the bottom, with the main flow of information going downward. You can either apply an automatic layout provided by the software tool, or create the layout manually. A manual layout may be time-consuming, but the result will highlight your message better (Tip 1). Often, a combination of a an automatic layout with a manual enhancement leads to the most satisfying results.

Our eye catches and our brain focuses on things we find attractive. Therefore, your map should be visually appealing to your audience if you expect others to look into the details of your work. Although such graphical characteristics are neither covered nor regulated by the SBGN, add carefully chosen colors, adapt the label fonts and the size of the symbol to the components, and keep arc lengths short, etc. To limit the number of crossing arcs, make use of clone markers to duplicate “hub nodes.” This will help break up the pathway and clear the visualisation, as shown in Fig 3. Beautify your network so people want to look at it!

When satisfied with the map you designed, save it in different formats. To be able to edit the map in the future, save it in the richest source format recognised by the software (e.g., Inkscape Scalable Vector Graphics (SVG) for Inkscape, or SBML with CellDesigner extension for CellDesigner). To ensure that your readers can view your map as you designed it, a raster file format (an image made up of pixels) such as Portable Network Graphics (PNG) is probably the best, or a vector format such as Portable Document Format (PDF) for a zoomable image. Finally, save the formal representation of your map in the SBGN Markup Language (SBGN-ML) format to increase compatibility between editor and viewer software (see S1 Supporting information for the SBGN-ML files of the figures). SBGN-ML is an Extensible Markup Language (XML)-based format specifically designed to store and exchange SBGN maps [3]. By providing your map in several formats, your work will be shareable and reusable by machines and by humans.

When publishing your data, provide the SBGN visualisations saved in formats as described in Tip 8, and link them to your original data, whether the data is a model, a publication, a data set, or other. It will benefit your work and help the readers to understand your network. You can, for instance, build a Computational Modeling in Biology Network (COMBINE) archive, a container format that facilitates the sharing and reproducibility of Systems Biology projects [14]. By using the web-based CombineArchive tool [15], you can easily and quickly create a COMBINE archive by uploading and packing all files relevant to the study, among which the newly generated SBGN file.

The SBGN editorial board provides and maintains guidelines, tutorials, and examples on their homepage (see S1 Table for a list of useful links). Feedback can be requested and questions can be asked on the SBGN mailing lists for end-users and developers (http://sbgn.github.io/sbgn/contact). The SBGN community is an extremely diverse and well-established community of computational biologists, modelers, computer scientists, and scientists from related fields. There will always be people willing to help, and questions are in general answered quickly. Remember, if you asked yourself a question, someone probably wondered the same before, so don’t be shy!