A western blot integrity checklist starts at acquisition, before any adjustment

Transformation and background subtraction are described as reasonable, and often scientifically necessary, depending on how complex the sample is and how specific the signal ^[7]. The condition attached is consistency. The same adjustment has to reach every lane in the comparison, and it has to be stated in the figure legend or the methods ^[8].

A separate instruction concerns the numbers. Densitometry should be run on raw images inside the linear range of detection, and a macro should be validated against a known standard before it is applied across a set of images ^[12]. That is a measurement requirement, and a lab can fail it without ever opening an image editor.

The case for all of this is made as prudence. The article attributes the scrutiny to the retraction of "a number of papers" over image irregularities, without a count or a period ^[3][19]. It does not report what share of submissions get flagged at screening, or how labs that archive unprocessed files compare with labs that do not. What the guidance does assert is that access to raw images is increasingly demanded at submission, and that keeping the files is the most practical way to demonstrate integrity and to speed review ^[9]. None of that says whether documentation changes the rate at which bad figures are produced, only whether a lab can answer when asked.

Six software packages are named. They are not interchangeable ^[20]. ImageJ, from the National Institutes of Health, is described as widely accepted by journals; it preserves image data, quantifies signal intensity directly, and supports macros so identical processing runs across many images ^[11]. Fiji is the recommended starting point for first-time users ^[13]. GIMP is listed for cropping, documentation and annotation, and the guidance says it is not designed for scientific quantitation ^[14]. Python with scikit-image is offered to researchers who want full control of the pipeline ^[15]. EBImage sits in the Bioconductor Project and hands off to R for downstream statistics ^[16]. Image Lab is described as preserving the context of the data from acquisition through analysis in a secure traceable environment ^[17].

The macro is the cheapest part of the record, because a validated macro applied to a whole set is itself the evidence that every lane got the same treatment ^[11][12]. Storing the unprocessed files is a decision made before the experiment.

Quantitation also depends on the normalization strategy chosen to correct non-biological differences between test samples, which the article turns to next ^[18]. Western blotting is still one of the most widely used methods in life science research, on grounds of affordability, specificity and accessibility, and mass spectrometry has not displaced it ^[2].