Preparing your manuscript: guidelines for writing microscopy methods and figures

A new paper has caught your eye on Twitter or Pubmed based on the title. Next step? A quick look at the abstract. Still interesting? Let’s have a look at the figures. This is probably the most important element of the paper to attract the interest of the reader. If figures are easy to interpret and well documented, the reader will probably decide to take the time to read the whole paper. Therefore, clear and informative image-based figures are key in attracting readers to your paper.

Similarly, when trying to implement a new imaging technique or just replicating an experiment, the methods section of the publication is crucial. Yet, sometimes this section is limited to: ‘image acquisition was performed on a [Manufacturer] confocal microscope’ or ‘image analysis was performed using ImageJ’. Here, clear and detailed reporting of image acquisition as well as bioimage analysis will ensure data reproducibility.

For these reasons, we thought it would be helpful to write a post with some guidelines to help you write your methods section and prepare your figures. Coincidentally, there have been some recent publications working towards the same goal that cover these aspects in detail [1-3]. We’ll be summarizing these publications here, but we highly recommend that you read these papers for a more exhaustive discussion.

Preparing clear and informative image-based figures

In a paper recently published in PLoSBiology, Tracey L. Weissgerber and colleagues [3] examined the legibility and interpretability of images in a total of 580 papers. The authors looked for common problems such as missing scale bars, poorly marked insets, accessibility for colour-blind readers, or insufficient explanations about images. Surprisingly, only 12 and 16% of papers examined in the fields of cell biology and physiology, respectively, met all good-practice criteria. This percentage went down to only 2% for plant sciences!

Below, we list standard practices for figure preparation as suggested in the Weissgerber paper (items 1-6). We also include other points worth considering (items 7-9). Aside from these suggestions, remember that many journals have their own guidelines and recommendations, which should be checked before submission.

1. Image magnification: should match the scientific question. Insets help to show more than one scale magnification.
2. Scale bar: always include a scale bar in your images. Make sure it is clearly visible (e.g. white scale bar on black background, black scale bar on white background). Include bar dimensions on the image, if possible, to facilitate the reader’s visualization. Also, ensure the thickness/size is still readable in the final printed version. Always check the proofs for readability.
3. Image colours: greyscale images have higher contrast. For multichannel images, it is recommended to include the single channel images in greyscale and the merged channels image in colour. It is also highly recommended to use colourblind-safe colours, for example, using magenta instead of red. This plugin on ImageJ can be useful to test your colour combination: https://biii.eu/simulate-color-blindness.
4. Follow a logical layout when designing multipanel figures: organise your panels in rows or columns to facilitate reading from top to bottom or from left to right. Ensure enough space between panels.
5. Image annotation: use arrows, stars, outlines or colour highlights to draw attention to details in the image, and letters or symbols to define features.
6. Figure legends: present information in a clear and accessible manner. Remember to include sample description and explanation of labels, annotations, colours, etc. If the images had any manipulations such as smoothing, denoising or any other filtering, specify in the figure legend.
7. Representative images: avoid choosing extreme phenotypes. If image analysis was performed, select an image that falls within the average measurement/observation.
8. Image manipulation or ‘beautification’: any adjustments to the entire image such as brightness, contrast, or smoothing filter should be detailed in the figure legend. If comparing images (e.g. control vs treatment), adjustments should be identical. Find more information about digital imaging ethics here.
9. Image quantification: avoid qualitative interpretation of imaging data and always try to report quantitative data interpretation.

Christopher Schmied and Helena Klara Jambor also wrote a very useful workflow (and cheat sheets) for making your figures using Fiji [4]. Of note, most of this section can also be applied to image-based slides for presentations.

Writing an imaging methods section

Recently, Teng-Leong Chew and colleagues [1] wrote a very detailed review on how to accurately report microscopy imaging parameters to ensure data reproducibility. The different components in the light path of the microscope such as light source, filters, objectives and detector contribute to the resulting image. Therefore, it is important to state all these parameters in your methods section to ensure reproducibility. If you don’t know what these parameters mean or how they can affect the image, please refer to the review by Chew and colleagues for some experimental examples. Below, we list the parameters you should include in your methods section depending on the imaging modality you used.

Widefield fluorescence microscopy:

• Microscope manufacturer
• Light source: mercury lamp, LED, etc.
• Objective lens: manufacturer, magnification, numerical aperture and optical aberration corrections
• Excitation/emission optics (filter cube): dichroic mirror (manufacturer and wavelength), excitation and emission filters (manufacturer and wavelength)
• Detector: type of detector, manufacturer, exposure time, gain, offset, binning
• Z-stack imaging: z-step size, total volume/number of steps
• Live-cell imaging: number of cycles, total duration time, time interval between cycles

Example: ‘Samples were imaged on a [Manufacturer] widefield microscope, equipped with a mercury lamp (manufacturer), using a [Manufacturer] Plan-Apochromat 63x/1.4 NA Oil lens. GFP was detected using a 520/36nm excitation filter(manufacturer, reference), 472/30nm emission filter(manufacturer, reference), and 495nm dichroic mirror (manufacturer, reference). The samples were illuminated at x mW power at the sample and acquired with a CCD camera (Manufacturer). Exposure time and camera gain were 100ms and 50, respectively.’

Confocal fluorescence microscopy:

• Microscope manufacturer
• Light source: laser wavelengths
• Objective lens: manufacturer, magnification, numerical aperture and optical aberration corrections
• Pinhole size
• Emission bandpass detection
• Detector: type of detector (PMT, GaAsP,…), dwell time, gain, offset
• Line/frame scanning accumulation or averaging (if any)
• Z-stack imaging: z-step size, total volume/number of steps
• Live-cell imaging: number of cycles, total duration time, time interval between cycles

Example: ‘Samples were imaged on a [Manufacturer] laser scanning confocal microscope, equipped with an Argon laser and a He/Ne 633nm laser, using a [Manufacturer] Plan-Apochromat 63x/1.4 NA Oil lens. Pinhole size was set to 1AU. Samples were illuminated with 488 and 633nm lasers at x and x µW, respectively. 1024 x 1024 pixel images were acquired using PMTs detectors. Alexa Fluor 488 acquisition parameters were: 500-600nm (emission wavelength range), 0.33µs (dwell time), 780 (gain) and 50 (offset). Alexa Fluor 633 acquisition parameters were: 640-750 nm (emission wavelength range), 0.33µs (dwell time), 800 (gain) and 0 (offset). Line average of 2 was applied to both channels. For z-stack imaging, a total of 10 slices were acquired with a 130nm z-step size.’

On this section, we focused on the two most commonly used techniques: widefield microscopy and confocal microscopy. Nevertheless, the imaging parameters for other microscopy techniques do not differ much from these ones. In general, you will find all these parameters in your image metadata. If you are not sure, ask the imaging facility personnel for help.

Writing a bioimage analysis methods section

The second paper by Chew and colleagues [2] focusses on how to properly report image processing methods. Given the importance of reporting quantitative data interpretation, as we discussed before, image processing and image analysis become another essential piece of the methods section.

As the authors point out in the paper: “…digitally processing raw images for later analysis is not intrinsically unethical; indeed, good use of processing techniques can be integral to achieving the experimental goals. What becomes unacceptable, usually unintentionally, is the failure to document the processing steps taken.”

Below, we list the parameters that the authors state should be reported (items 1-5) and some other points we think are also important to consider (items 6-8). If you are not sure what each of these parameters mean or how they can affect your image, please refer to the Chew and colleagues review for clear and comprehensive examples.

1. Background subtraction: report algorithm used (e.g Gaussian smoothing) and parameters (e.g. kernel size and shape)
2. Denoising: report algorithm and parameters
3. Deconvolution: report commercial software (if used), algorithm and parameters (e.g. PSF and number of iterations)
4. Intensity threshold: if manual, report intensity value or percentile used. If automatic, report which method was used (e.g. Otsu, triangle).
5. Segmentation: report all parameters used for segmentation such as pixel connectivity, morphometric filtering, and binary operations
6. Analysis: report any software, script or plugin used for image quantification
7. References: remember to include the references of any open source software (yes, this includes Fiji) or plugin used. The developers of these tools also depend on citations and funding for their research.
8. Share your code: include any workflow macros (e.g. ImageJ macro) and other custom software codes as supplemental material or share them via GitHub or public data repositories such as OMERO or Zenodo

Again, if your imaging facility or a bioimage analyst helped you with the imaging analysis, ask them for help to prepare this methods section.

To conclude: acknowledgments section. If you used a core facility for your imaging experiments or image analysis, don’t forget to acknowledge them. This is an important metric to evaluating and funding core facilities. In the end, your research will benefit from this, too. Consider even adding personnel as coauthors if they made significant contributions to the paper.

Since there’s a lot of information here to remember, we made a checklist document listing these standard practices and parameters. If you think we missed anything, please let us know. We hope this helps you in the preparation of your manuscript.

References

1. Heddleston JM, Aaron JS, Khuon S, Chew TL. ‘A guide to accurate reporting in digital image acquisition – can anyone replicate your microscopy data?‘ J Cell Sci. 2021 Mar 30;134(6):jcs254144. doi: 10.1242/jcs.254144.

2. Aaron J, Chew TL. ‘A guide to accurate reporting in digital image processing – can anyone reproduce your quantitative analysis?‘ J Cell Sci. 2021 Mar 30;134(6):jcs254151. doi: 10.1242/jcs.254151.

3. Jambor H, Antonietti A, Alicea B, Audisio TL, Auer S, Bhardwaj V, Burgess SJ, Ferling I, Gazda MA, Hoeppner LH, Ilangovan V, Lo H, Olson M, Mohamed SY, Sarabipour S, Varma A, Walavalkar K, Wissink EM, Weissgerber TL. ‘Creating clear and informative image-based figures for scientific publications‘ PLoS Biol. 2021 Mar 31;19(3):e3001161. doi: 10.1371/journal.pbio.3001161.

4. Schmied C and Jambor HK. Effective image visualization for publications – a workflow using open access tools and concepts [version 2; peer review: 2 approved]. F1000Research 2021, 9:1373 (https://doi.org/10.12688/f1000research.27140.2).

(3 votes, average: 1.00 out of 1)