Can Combining Images From Different Experiments Create a Misleading Research Figure?

Can You Assemble One Research Figure From Multiple Images?

Research figures are often composites. A multipanel figure may contain microscopy images from several conditions. Nonadjacent gel lanes may be placed next to each other. Representative images may come from different samples. Sometimes researchers need to combine visual material simply to communicate an experiment efficiently.

The existence of a composite is therefore not inherently suspicious.

The problem begins when the finished figure creates a false impression about where its components came from or how they relate to one another. Readers should not be led to believe that images came from the same experiment, field, gel, exposure, sample, or experimental context when they did not.

A Composite Is Acceptable Only If Its Provenance Remains Truthful

The Scientific Problem Is Not the Composite Itself but What the Composite Implies

Many Legitimate Figures Combine Multiple Images

A figure containing separate panels labeled A, B, C, and D may legitimately show images from different experimental conditions, samples, participants, time points, or replicates. Readers generally understand that clearly separated panels can have different origins.

Likewise, researchers may sometimes need to juxtapose nonadjacent regions of an image or gel for efficient presentation, provided the assembly follows applicable standards and does not disguise the underlying relationships.

The relevant question is therefore not “Were multiple files used?” It is “What does the final figure tell the reader about those files?”

A Composite Becomes Misleading When Separate Origins Are Made to Look Continuous

Suppose two microscopy fields are cropped and joined seamlessly so that the final figure appears to show one continuous field of view. Unless the composite nature is clear, a reader may reasonably assume that all visible structures occupied the same spatial context.

That assumption could matter scientifically. Spatial relationships, cell density, proximity, distribution, and local context may all contribute to interpretation.

Joining separate images while concealing the boundary can therefore create information that was never actually observed in one field.

Gel and Blot Splicing Requires Particular Care

Researchers sometimes rearrange lanes from electrophoretic gels or blots. Current Nature Portfolio image-integrity guidance, for example, states that rearranged nonadjacent lanes must be clearly delineated and that the rearrangement should be stated in the figure legend.

The reason is straightforward. Adjacency can imply that samples were next to one another on the original gel and shared the same experimental and imaging context.

More serious problems arise when bands or lanes from unrelated experiments are assembled into what appears to be one coherent experiment. ORI has made research misconduct findings involving Western blot panels created by splicing material from separate unrelated experiments and presenting the composite as though it represented another experimental context.

“Same Experiment” and “Different Experiment” Can Matter for Comparison

Images produced under different experimental runs may have different acquisition conditions, exposures, reagent batches, instrument settings, backgrounds, or other sources of variation.

Putting them side by side can still be scientifically informative. But a figure should not imply that differences between them necessarily reflect the experimental variable of interest when run-to-run variation could also contribute.

Nature Portfolio guidance specifically cautions against quantitative comparisons between samples on different gels or blots and requires disclosure when such comparisons are unavoidable and derive from the same or parallel experiments processed in parallel.

Separate Panels Do Not Need to Pretend They Share One Physical Image

One of the easiest ways to avoid confusion is simply to preserve panel boundaries.

If a control micrograph and treatment micrograph come from separate fields, presenting them as two clearly labeled panels tells readers what they are. There is usually no scientific advantage in blending their borders until they appear to form one continuous scene.

Good figure design can be visually clean without erasing provenance.

Reusing the Same Image as Though It Represents Another Experiment Is Different

Combining images should not be confused with duplication and relabeling.

If the same microscopy image is presented once as “Control” and again as “Treatment,” readers are being shown one observation as though it were evidence from two different experimental conditions.

ORI has repeatedly reported misconduct findings involving images that were duplicated, manipulated, and relabeled to falsely represent different experiments or conditions. In one recent case, microscopy images were duplicated, flipped, lightened, or otherwise altered and presented as results from different experimental conditions.

That is not merely a formatting choice. It changes the claimed provenance of the evidence.

Changing Orientation Does Not Create a New Observation

Flipping, rotating, cropping, recoloring, or changing the contrast of an image does not transform it into independent evidence.

If the same source image is reused under another condition after rotation or reflection, it remains the same source image. ORI findings provide numerous examples in which duplicated images were flipped or otherwise manipulated before being relabeled as different experimental results.

Researchers should therefore maintain source-image identifiers or another provenance system that makes accidental reuse easier to detect during figure preparation.

Images From Parallel Experiments Need Accurate Labels

Suppose researchers perform the same experiment three times. A control image from replicate 1 and treatment image from replicate 2 happen to be the clearest examples.

Whether placing those images together is appropriate depends on what the figure claims to show, the research method, and the relevant reporting standards. If the panels are presented merely as examples of the respective conditions, their different replicate origins may or may not require specific disclosure depending on context.

But the figure should not create a false impression that the two images form a directly matched pair if they do not.

Controls Require Special Attention

Controls provide the baseline against which experimental evidence is interpreted. Reusing a control from another experiment without appropriate justification or disclosure can therefore create a false comparison.

For gels and blots, some publisher policies impose explicit requirements. Nature Portfolio, for example, states that loading controls must be run on the same blot and that sample-processing controls run on different gels must be identified as such in the figure legend.

The applicable rule may differ for other image types, but the underlying principle is broader: readers should be able to understand what was actually compared.

Figure Legends Cannot Rescue a Fundamentally False Composite

Disclosure is important, but it is not magic.

If combining images creates a scientifically invalid comparison, adding “images combined from separate experiments” to the legend does not necessarily make the comparison methodologically sound. Transparency allows readers to evaluate what was done; it does not convert an invalid procedure into a valid one.

Researchers need both an appropriate scientific rationale and accurate presentation.

Composite Figures Should Remain Traceable to Their Source Files

For each panel, lane, or image region, researchers should be able to determine the corresponding original data.

This can be achieved through sensible file naming, figure-generation scripts, laboratory records, metadata, source-data files, figure maps, or other methods appropriate to the project.

Traceability becomes especially important when figures contain many similar images. Without it, accidental duplication and mislabeling become surprisingly easy. A figure assembled at 2 a.m. before a submission deadline remains subject to the same laws of provenance as one assembled after coffee.

A Composite Can Be Misleading Even If Every Component Is Authentic

This is the most important conceptual point.

Imagine that every image in a figure is genuine and completely unedited. The figure can still misrepresent the research if those authentic images are labeled as different conditions, assembled to imply a shared experimental context, or juxtaposed in a way that creates a comparison the underlying experiments do not support.

Pixel authenticity and figure accuracy are related but distinct.

Misleading Composites Can Fall Within Falsification

The PHS definition of falsification includes manipulating research materials, equipment, or processes, or changing or omitting data or results such that the research is not accurately represented in the research record.

ORI has made misconduct findings involving composite images assembled from unrelated experiments and falsely represented as other experimental results. For example, an ORI case involved Western blot panels spliced from separate unrelated experiments on different cell lines and presented as a composite representing another source.

Still, finding an unexplained composite does not by itself prove misconduct. Original files, figure histories, accepted practices, and evidence about how and why the figure was assembled remain necessary for evaluating a particular case.

Four Gel Lanes and Two Very Different Figures

Common Misunderstandings About Composite Research Images

Build Composite Figures So Their Experimental History Remains Visible

When assembling a figure, think of each panel or image region as carrying provenance. You should be able to trace it backward to the source file and explain why it appears where it does.

Keep the original files and enough figure-construction information to reconstruct the assembly. That contributes directly to the broader task of preserving evidence that research images are authentic.

Also remember that assembling the images correctly does not settle whether they were chosen fairly. A technically flawless composite can still mislead if the researcher selects only images that support the preferred conclusion.

Before Publishing a Composite Research Figure, Check This

Frequently Asked Questions About Combining Research Images

A Composite Figure Should Combine Images Without Combining Their Histories

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.