What is Dr. Zubair Khalid's research focus?

Dr. Zubair Khalid specializes in molecular virology, mRNA vaccine development, and computational biology, with a focus on avian pathogens like IBDV and Avian Reovirus.

Where is Dr. Zubair Khalid currently working?

Dr. Zubair Khalid is a Postdoctoral Research Associate at the University of Maryland (UMD), specifically within the Department of Animal and Avian Sciences.

Research Figures: Making Methods and Results Easier to Evaluate

A well designed figure is the fastest way for a reader to understand your methods and results. This guide covers the core decisions that determine whether a figure informs or misleads. It is written for graduate students, postdocs, and principal investigators who prepare figures for manuscripts, grants, or presentations. Every section draws on established practices from the NIH Office of Intramural Training and Education and other authoritative resources [1]. The goal is not to dictate a single style but to give you criteria you can apply to your own work.

If you share your raw data alongside a figure, reviewers can verify your analytical choices. The NIH Data Management and Sharing Policy now requires many researchers to plan for data access [4]. A figure that openly refers to its source data is easier to evaluate than one that hides it. This guide assumes you want to build figures that stand on their own and pass the scrutiny of peers.

At a Glance

Element	Best Practice	Why It Matters
Panel logic	Arrange panels to tell a story from left to right, top to bottom	Reduces cognitive load and guides the reader through your narrative
Scales	Start axes at zero when possible, avoid truncation that exaggerates differences	Prevents visual distortion of effect sizes
Color	Use a limited, perceptually uniform palette, avoid red green combinations	Improves clarity and supports color vision deficiency
Annotations	Label key features directly on the image, use consistent font size	Makes the figure self contained without forcing readers to flip to the legend
Source data	Provide raw values or a link to the repository	Enables reproducibility and trust in your conclusions
Accessibility	Use text alternatives and high contrast elements	Broadens your audience and meets funder expectations
Misleading choices	Do not cherry pick time points, compress outliers, or omit error bars	Preserves scientific integrity and reduces misinterpretation

The Logic of Panel Arrangement

Each figure should have a clear beginning and end. Readers scan panels in a Z shape if your language reads left to right. Place controls or baseline conditions in the first panel and experimental outcomes in the last. For example, a recent study on copper metalation in hydrogen bonded organic frameworks used a carefully sequenced set of panels to show synthesis, characterization, and device performance [5]. That sequence made it easy to follow the logic of the paper. When you design your own figure, ask yourself: can a colleague who skips the text still understand the main point by reading only the panels? [1]

Group related panels together. A dose response curve belongs next to a bar chart of the same data, not separated by a microscopy image. Use white space or thin lines to separate distinct experiments but keep them within the same figure when they support a single conclusion. Avoid cramming twelve panels into one figure if three well designed ones would serve better.

Setting Scales and Axes

The scale you choose changes the visual story. A bar chart that starts at a number other than zero can make a tiny difference look huge. For continuous variables, always start the y axis at zero unless you have a strong statistical reason to do otherwise. If you need to show a magnified view, add a break symbol or use a second panel. The Chi Square test for independence relies on expected frequencies that are not distorted by scaling artifacts [7]. That principle applies equally to figure design: do not let axis scaling distort the viewer's perception.

Log scales are appropriate for data that span several orders of magnitude. Label the axis clearly and include the base of the logarithm (log10, ln). Avoid mixing linear and log scales in the same figure unless you add a clear annotation. A systematic review of intrapleural fibrinolytic therapy in children included forest plots with natural logarithmic scales for odds ratios, and the authors carefully labeled each axis to prevent confusion [8]. That attention to scale made the results immediately interpretable.

Choosing and Using Color

Color should encode information, not decoration. Use a sequential palette for ordered data (low to high) and a diverging palette for data that fall around a meaningful midpoint. Avoid pure red and pure green together because about 8 percent of men have some form of color vision deficiency. Tools like ColorBrewer or viridis provide perceptually uniform options. A study on physical activity and cognition in older adults used a color coded map to show regional differences, and the authors selected a palette that remained distinguishable when printed in grayscale [9]. That choice respected both accessibility and practicality.

Limit your palette to three to five distinct colors. More than that forces the reader to constantly check the legend. Use color only for the variable you want to emphasize. Gray or black can carry structural elements like axes and gridlines. If you use color for statistical significance (for example, asterisks in red), state that convention in the legend.

Annotations and Labels

Every figure element needs a label. Panels should have a letter (A, B, C) in the top left corner. Axes need titles with units in parentheses. Directly annotate significant features on the image itself. For a micrograph, draw arrows to the structure you want to highlight and add a short description. A recent study on GABARAPL2 and Alix regulation included fluorescence images with arrowheads pointing to colocalized puncta, and a one line label explained each condition [6]. That approach is far more helpful than a long legend that forces readers to search.

Choose a legible font, typically 8 to 12 points for print or a proportional size for presentations. Use the same font everywhere in the figure. Bold only the labels that need emphasis, such as panel letters or key annotations. Check that all text remains readable after the figure is shrunk to fit a journal column.

Providing Source Data and Reproducibility

A figure without source data is an assertion, not evidence. Journals increasingly require a data availability statement or a supplementary file that contains the raw values used to generate each plot. The NIH Data Management and Sharing Policy encourages researchers to deposit data in recognized repositories [4]. You can also overlay individual data points on bar charts or box plots. This practice, sometimes called showing the beeswarm, reveals distribution shapes and outliers that summary statistics hide.

Include the sample size (n) for each group directly on the figure or in the legend. If you performed statistical tests, report the test name, the test statistic, and the p value either on the figure or in a parenthetical note. A meta analysis of thoracic imaging for empyema management used forest plots that displayed each study's weight and confidence interval, making the evidence transparent [10]. Your figures should aspire to that same transparency.

Accessibility Considerations

Figures are inaccessible when they rely solely on color to convey meaning. Add patterns, shapes, or text labels to distinguish groups. Ensure that all text has sufficient contrast against the background. The NIH Office of Intramural Training and Education provides resources on creating accessible scientific content [1]. Many funders now require a caption that describes the figure in words. Write a short descriptive paragraph that gives the overall finding before listing details.

If you use a graphing software, check its output for accessibility features. Avoid laser printed figures that use thin lines, because small variations disappear under photocopying or projection. For online publication, provide a high resolution PNG or SVG that can be zoomed without pixelation.

Common Misleading Choices and How to Avoid Them

Cherry picked time points. If your experiment runs for 48 hours, do not show only the 6 hour point unless you have a pre registered reason. Show the full time course in a supplementary figure if space is tight.

Truncated axes. Cutting the y axis to amplify differences is a classic manipulation. Always consider whether a reader would interpret the figure differently with a zero origin.

Omitting error bars. A mean without some measure of variability is incomplete. Use standard deviation, standard error, or confidence intervals, and state which one is shown.

Overlapping data points. When many points land on top of each other, the figure obscures density. Use transparency, jittering, or a contour plot instead.

Mismatched scales in paired panels. If you compare two conditions, keep the scale the same unless you explicitly state otherwise. Changing scales between panels invites accusation of bias.

Limits and Uncertainty

No figure format is universally correct. Conventions vary by field. Cell biologists may need high resolution micrographs with scale bars, while epidemiologists use Kaplan Meier curves with at risk tables. This guide emphasizes principles that transfer across disciplines, but you must adapt them to your audience.

A figure cannot replace a thorough methods section. Some details, such as antibody dilutions or statistical models, belong in the text even if a figure is self explanatory. Also note that journals have specific requirements for resolution, color mode, and file format. Always check the author guidelines before finalizing.

Frequently Asked Questions

How many panels should a single figure contain? Aim for three to six panels that tell one coherent story. If you have more than eight panels, consider splitting the figure into two or moving some to supplementary material.

Can I use red and green if I note that they are colorblind safe? No. A note does not help a colorblind reader distinguish the two values. Use textures or shapes instead, or choose a palette that is distinguishable without color.

Should I always show individual data points? Not always, but you should show the distribution. For small sample sizes (less than 20 per group), individual points are strongly recommended. For large datasets, a violin plot with median line works well.

What is the best file format for figures? Vector formats (SVG, EPS, PDF) are preferred for graphs and diagrams because they scale without losing quality. For micrographs, use a high resolution TIFF or PNG (300 dpi minimum).

Zubair Khalid