A row of dark smudges on a strip of membrane, each one a protein caught mid-journey through a gel. That is a Western blot, and it is the workhorse image of cell biology. It is also, on this occasion, the thing that went wrong. In a correction published on 29 July 2026, the Journal of Oncology records that in a 2012 paper on ovarian cancer cell lines, the bands in the fourth panel of Figure 1a were "incorrectly oriented during figure assembly."
The original study, by S. M. Moghaddam, A. Amini, A.-Q. Wei, M. H. Pourgholami and D. L. Morris, looked at two proteins called Sprouty-1 and Sprouty-2 across a panel of human ovarian cancer cell lines: OVCAR-3, SKOV-3, 1A9, A2780, CAOV-3, OV-90 and IGROV-1, compared against HOSEpiC, a line of normal human ovarian surface epithelial cells. The point of such a comparison is to see whether a protein is present at different levels in cancerous cells than in healthy ones.
The panel at issue did not show Sprouty at all. It showed GAPDH, an abundant housekeeping protein that researchers use as a loading control: a way of confirming that roughly equal amounts of protein were put into each lane of the gel, so that a faint band elsewhere means "less of this protein" rather than "less of everything." Get the loading control wrong and every comparison above it becomes harder to trust. The corrected figure caption now also states plainly that the GAPDH loading control bands were reused for both the Sprouty-1 and the Sprouty-2 analysis.
What the authors added
Along with the fixed figure, the authors volunteered a longer explanation of how the experiment ran, aimed at readers trying to interpret the image. It is unusually specific, and it addresses an obvious question: if Sprouty-1 and Sprouty-2 travel through a gel at roughly 34 to 36 kilodaltons, and GAPDH travels at roughly 36 to 37, how did anyone tell them apart?
The answer, the authors write, lies in the gel itself. They resolved the proteins on 10 to 12 percent SDS-PAGE gels, a formulation that separates proteins in the 30 to 40 kilodalton range and can distinguish differences of 1 to 3 kilodaltons. The proteins sit in a similar weight range, they acknowledge, but under those conditions they are distinguishable.
The blot came from a single PVDF membrane, probed in sequence. After detecting the first target, the team stripped the membrane and reprobed it for the next, ending with GAPDH. Sprouty-1 and Sprouty-2 were detected with mouse monoclonal antibodies from Abnova; GAPDH with a monoclonal antibody from Sigma-Aldrich. Antibody specificity, the authors argue, provides a second check on which band is which.
They are also candid about what is missing. Molecular weight markers, the reference ladder that lets a reader eyeball how far a band has traveled, were run during the original electrophoresis. But the published figure showed cropped panels and left the marker lanes out, which the authors describe as common practice in many publications at the time. Band identity therefore rested on three things: expected migration position, antibody specificity, and reproducibility across independent replicates run during the study. The correction closes with four words: "We apologize for this error."
Why it matters
A correction notice is not a retraction. Nothing here says the original conclusions about Sprouty proteins in ovarian cancer are wrong, and the notice makes no claim in either direction about the biology. What it does is repair the visual record and put on the page several methodological details that the 2012 figure did not carry.
That second part is the more interesting one for a general reader. Cropped Western blots without marker lanes were, by the authors' own account, ordinary in the literature not long ago. The expectation has shifted. Journals and readers now want to see the seams: which lanes were reused, whether a membrane was stripped and reprobed, how confident anyone can be that a band at 35 kilodaltons is the protein claimed rather than its close neighbor. This notice is a small example of an older paper being brought up to the newer standard, fourteen years after the fact.
It is worth being clear about the limits of what a document like this can tell you. The correction reports what the authors say happened at the bench in or before 2012; it does not present new experiments, new images beyond the corrected panel, or any independent verification. Readers wanting to know whether Sprouty-1 and Sprouty-2 really do behave differently in ovarian cancer cells will still need to go to the original paper, and to whatever work has followed it.