Writing a scientific paper? Preparing for publication?
Making a poster for a conference tomorrow? Let’s face it: People will skim through the body text of your paper/poster, and only look at the figures. We all love figures. They catch eyes, they deliver information effectively, they highlight your results.
Figures are the star of your paper. Your job is to make them shine.We’ve talked about good and bad charts before. After making a good chart (or a figure of any kind), you will have to describe it. That’s what the legend (a.k.a caption) is for. A good legend will make or break a graphic.
But if the result is too easy to reproduce by a non-enthusiastic reader in half an hour, it is a bad scientific paper. Outrageous hypotheses. We all need to form a hypothesis to be proven of disproven later by the research steps, or at least most of us do. When the editors of any journal are doing a good job (focused on science), the results are good and can overcome bad reviewers (who will always be there). It seems like the editors at SciRep are more focused on the meaningless than the quality of what is published. PS You didn’t tell us there was math involved in posting a comment!
This post will give you a few tips to write and format figure legends for scientific papers.Write it right: 3 notes about the content of a legendA figure should be able to stand alone. Readers want to understand it without going back and forth between the figure and the text sections.
Don’t tell your readers to “See Methods” — most of them won’t look through the maze of text to search for the exact spot where you buried this piece of information.How to write a good legend that makes a figure clear as day, all by itself, in a scientific paper? The title:The title should tell the readers “This is what this figure is about” very clearly. Use active voice, and keep it short.Make it a sentence that summarize the major result seen in the figure.Example: Hippocampal neurons derived from patients with bipolar disorder show hyperexcitability.Or, make it a phrase stating the type of analysis used.Example: Quantification of XYZ transgene expression using RT-PCR. 2.The methods:This is where things tend to get out of hand. It is true that you should include lots of details, like assay name, antibodies used, cell type or animal model, treatments and controls, statistical tests, numbers of replicates, et cetera. The figure must be able to stand alone.
You want your readers to understand it without going back to the method section.But on the other hand, you don’t have to rewrite the method section. Limit what to include in the legend to the absolute minimum that is required to understand the figure.Example: Average expression of representative genes involved in the PKA/PKC and AP firing systems revealed by RNA-seq (a) and qRT–PCR (b) analysis (normal, n=4; BD, n=6 lines). The results:You really only need this if i) the title did not state it already, or ii) the figure has multiple panels or shows multiple results. Either way, contain the temptation to go on and on about your findings, and keep this part short and sweet.
Limit it to one or a few sentences that describe the key findings that are seen in the figure.Example: Patch-clamp recording on Prox1::eGFP-expressing neurons showed spontaneous postsynaptic currents.Side note: You might not have a choice to keep the legends short. If you are publishing with Science or Nature (congratulations!), they put a limit on body text, so a big chunk of the Methods and Results sections ends up in figure legends, hence their “signature” super long legends. A rule of thumb: Always follow the journal’s specific instructions. Figure legend example. Pretty please: 3 notes about the format of a legendMaking your figure legends look good doesn’t just satisfy your inner artist. An attractive, professional looking legend grabs attention. Here’s a few keys to get “the look.” 1.
Avoid clutters:When possible, place all labels on the graphics. That will make it easier to read both the graphs and the legends.Tone down on abbreviations. If you must use them, make sure they are consistent with the text of the paper, and that they are common acronyms and not the obscure system that only your lab uses to manage the tubes. Pick a good fontThe debate over the best font for a graph legend in scientific papers has not come to an end.
Meanwhile, here’s the consensus so far:. Follow the journal’s instructions (duh!). If there is no instruction, use the same font as the body text, and run it one size smaller. DO NOT TYPE ALL CAPS!Still overwhelmed by options?
![Have Have](http://abacus.bates.edu/~ganderso/biology/resources/writing/graphparts2003.gif)
Pick one from the “People’s Choice” fonts: Arial, Helvetica, Computer Modern, or Times New Roman.
Here are some recommendations for making scientific graphics which help your audience understand your data as easily as possible. Your graphics should be striking, readily understandable, should avoid distorting the data (unless you really mean to), and be safe for those who are colourblind. Remember, there are no really “right” or “wrong” palettes (OK, maybe a few wrong ones), but studying a few simple rules and examples will help you communicate only what you intend. What kind of palettes for maps?For maps of quantitative data that has an order, use an ordered palette.
If data is sequential and is continually increasing or decreasing then use a brightness ramp (e.g. Light to dark shades of grey, blue or red) or a hue ramp (e.g. Cycling from light yellow to dark blue). In general, people interpret darker colours as representing “ more”. These colour palettes can be downloaded from. A example categorical map from the US census of 2000.
AccessibilityTry to ensure that your palette is accessible to those with colourblindness. Especially avoid using red and green together, as these are difficult to distinguish for those with the most common forms of colourblindness – about 8% of the male population and 0.5% of the female population. The diagram below from, shows a simulation of a severe form of colourblindness (deuteranopia). There are a number of and that allow you to check how your visualisations will appear to those with colour blindness. If in doubt, remember that red and green should never be seen – at least not together.
A simulation of severe colourblindness.There is a great example of a colourblind friendly categorical palette in this.You can get this palette into a programming language using the RGB or CMYK values, or with the hex codes listed below. For example, you can create a colour palette in the programming language R, with the simple command: cbPal. A spectral palette, versus one with better perceptual qualities.Sometimes, a visualisation can benefit from using a spectral palette to highlight an important part of the data, or to split the data into categories. Understanding when this is appropriate will help your viewer understand the data. Many colours have widely culturally agreed meanings: red is associated with hot, stop or danger, whereas blue is associated with cold, rain or wet. Choose culturally appropriate colours where possible.
End of the RainbowDon’t use a “rainbow” or similar palette if you can avoid it. A “rainbow” or “spectral” type palette.Rainbow palettes can be useful –. Rainbow palettes combine several of the problems mentioned above – they use both red and green, place yellow in the upper mid-range rather than the top, and introduce perceived sharp transitions in places where none exist in the data,. See for some great research on the problems with “rainbow” type palettes. You can see many of these problems in this map using a “rainbow” type palette, taken from a and spotted.
Although it looks as though there is a huge east/west step change in the centre of the USA, a look at the legend reveals that the transition is in fact quite smooth. This is a great post, very useful.On the rainbow palette, especially in reference to the precip map of the US. This could actually be really useful if the yellowish-green to green transition was set at the generally accepted threshold for growing a particular crop without irrigation. When I saw the map, the famous “corn line” is the first thing I thought of the cline west of which you can’t really grow corn (in traditional farming) without irrigation, or the division between Eastern forests and prairies of the plains, etc. Obviously this would be more appropriate using evapotranspiration as the variable. If the evapotranspiration limit for corn (or some other crop) was set right at that yellow/yellow-green margin then the map would have powerful meaning, in this case, intended.Now, here’s the technical question.
If you have a set of software that produces graphs (or gradients as in drawing software) how to you teach the software to use a specific appropriate palette? Shouldn’t graphic and drawing software have built in named palettes that have the appropriate characteristics? (I see some of this is in the links you provide, but it would be great to see a full treatment of this.). Hi Greg,Many scientific users will be able to swap palettes in and out pretty much at will, as they will be generating the code to plot their figures. Often, they can even create their own palettes.
There are examples in R, python, matlab etc. For commercial software that you interact with, rather than code yourself, there should be a good set of palette options in there somewhere, if the software is worth paying for ?My guess is that people’s data sets and needs might be very specific, and that it is probably worth teaching the underlying principles, rather than having a load of go-to palettes. There’s a palette that is often used in climate science that mixes the useful distinct categories of a rainbow palette with the clarity of a diverging palette: Have darkish blue on the left, shading to pale green just below zero; switch to pale yellow just above zero, then shade to darkish red. The distribution of both hues and luminances is much more balanced than on an HSV rainbow (and the greens and reds are distinguished by having luminance). I haven’t seen an automatic way of generating it yet, and it’s not quite any of the ColorBrewer palettes.