Research integrity
Image presentation for scholarly communications and grant applications
These guidelines are mostly taken from the EMBO Molecular Medicine Guidelines for Authors and are intended to complement or integrate specific indications from funding agencies and scholarly journals. The ultimate goal is to ensure data integrity and reliability and to protect OSR investigators from unfounded post-publication allegations of misconduct originating from non-malicious but poor image processing and handling.
Image-based data (gels, micrograph photographs, etc.) should be scanned or captured at the highest resolution possible and saved in lossless formats. This can be achieved by saving images as TIFF while ensuring that no compression option is selected in the application used for acquisition/export of the images. It should be noted that 300 dpi at print size is usually considered the lowest acceptable resolution for original images for most journals; therefore, the default settings on imagers, scanners, and cameras should always be checked. Screenshots should never be used to capture images. If image file size is a concern, it is best to use lossless image compression such as LZW and in any case, to avoid quality-degrading compression formats such as JPEG. This is important because failure to provide sufficiently high-quality images leads to delays in publication, also due to the impossibility to verify data integrity. Many issues related to image quality, resolution (e.g. compiled figure vs. image resolution), format and others are due to misunderstanding of these basic concepts. It is therefore suggested that authors familiarise themselves with them, and may wish to refer to detailed resources on the topic from EMBO Press and The Journal of Cell Biology.
In preparation for submission e.g. of a manuscript to a journal or a grant application to a funding agency, a certain degree of image processing is acceptable (and for some experiments, fields and techniques unavoidable), but should be minimal (for instance, to add arrows to a micrograph) and the final image must accurately reflect the original data and conform to community standards.
The corresponding author should retain unprocessed (source) data and metadata files and be prepared to promptly make them available, as editors or any authorised office/person may request them at any time. It should be noted that if source data are unavailable, further consideration of a submitted manuscript may be stalled until resolution of the issue or stopped altogether.
It is suggested that the authors list all image acquisition tools and image processing software packages used in the preparation of figures and store this information in the README file associated with the manuscript folder. Authors should also document, or be prepared to document upon request, key image-gathering settings and processing steps in the Methods or other section of the manuscript, depending on journal policy.
Images gathered at different times or from different locations should not be combined into a single image, unless it is stated that the resultant image is a product of time-averaged data or a time-lapse sequence. If juxtaposing images is nonetheless deemed essential, the borders should be clearly demarcated in the figure and described in the legend. Typical scenarios in this respect are the “splicing” in of western blot sections and combination of different image fields into a single image (see further below).
The use of touch-up tools, e.g. cloning and healing tools in Photoshop or any other tool that deliberately conceals manipulations, must be avoided.
Image processing (e.g. changing brightness and contrast) is to be used with restraint and should not cause data loss (see further below). Furthermore, it is appropriate only when applied equally across the entire image including controls. In fact, processing to emphasize one region in the image at the expense of others is inappropriate and can amount to data falsification.
The policies for image processing and handling in reputable journals are quite stringent and provide clear instructions on responsible conduct and disclosure when certain image manipulations are necessary. For instance, the display of cropped gels and blots in a manuscript is usually permitted when aimed at improving the clarity of the presentation. Nevertheless, even when journal policies may be more permissive, OSR investigators should respect the following general data integrity indications for image presentation.
- Cropped gels should preserve all important bands, and sufficient space (several band-widths) should be retained above and below the relevant band(s).
- Vertically spliced images that juxtapose lanes that were not originally adjacent must be declared and have a clear white space or a black line indicating the splices. Quantitative comparisons between samples on different gels/blots are discouraged; if this is unavoidable, the figure legend must reflect that the samples derive from the same experiment and that gels/blots were processed in parallel.
- Highly-contrasted or otherwise-filtered gels and blots are far from optimal, as these manipulations may hide data. Ideally, exposures with grey backgrounds are preferable. If, however, this is unavoidable and necessary to reveal otherwise difficult-to-see details, then multiple exposures should be provided.
Items (cells/colonies/portions of tissue/etc.) from different fields should not be collated into a single field. While the principle of showing representative fields remains valid, the authors should offer, or be prepared to show multiple representative fields.
As noted above, linear transformation adjustments such as brightness and contrast adjustments are acceptable only if:
- applied to the entire image,
- applied identically to all images within the same experiment (e.g. treatment vs. control)
- does not lead to under- or over-saturation with subsequent loss of information.
Threshold manipulation, expansion or contraction of signal ranges and the altering of high signals must be avoided. If ‘Pseudo-colouring’ and nonlinear adjustment (for example ‘gamma changes’) are used, this must be disclosed. It is appreciated that adjustments of individual colour channels are sometimes necessary on ‘merged’ images but, again, this should be disclosed in the figure legend.
Best practice on data integrity would suggest the inclusion of the following information in a scholarly publication:
- The type of equipment (microscopes/objective lenses, cameras, detectors, filter model and batch number).
- Acquisition software used.
- Equipment settings for critical measurements.
Authors should also have the following information readily available for each image:
- Acquisition information, including time and space resolution data (XYZT and pixel dimensions); image bit depth.
- Experimental conditions such as temperature and imaging medium.
- Fluorochromes (excitation and emission wavelengths or ranges, filters, dichroic beam splitters, if any).
- The display lookup table (LUT) and the quantitative map between the LUT and the bitmap, especially when rainbow pseudocolour is used. If the LUT is linear and covers the full range of the data, that should be stated.
- Measured resolution at which an image was acquired and any downstream processing or averaging that enhances the resolution of the image.