Lu et al. (2004b), Supp. Fig. 2E: An unexpected line and unexpectedly similar patterns in background and mRNA bands are visible.
Background: The authors reported that these Northern blot experiments provided “evidence for absence of Unc5b transcripts” in the novel Unc5b mutant that was generated for this study. These experiments were essential for validating the use of the Unc5b mutant line in other experiments reported in this study, and thus provided fundamental support for the authors’ conclusions.
Concerns:
Histogram analysis and PCA were performed using Forensically.
I believe the similarities noted in the +/+ lanes of two reported Northern blot results likely come because the same blot has been hybridized twice, but with the two different probes. This is a normal and accepted practice that makes it possible to compare hybridization results of different probes on the very same samples/blots. Note that the unc5b mutant was made by knocking Bgeo/PLAP reporter genes into the locus downstream of exons 1 and 2 (see gene diagrams in SuppFig2a and 2b). The full length probe (exons 1-17) can detect both the normal sized transcripts in +/+, and a larger transcript that comes from the knock in allele (in +/m and m/m lanes), as expected. In contrast, the exons5-17 probe is missing the unc5b early exons, can still detect the wild type transcripts, but no longer detects the larger sized transcript that is only expected to have early unc5b exons spliced into the the Bgeo/PLAP inserted fragment. I think the data presented support the authors' description of the engineered mutation. This is also a case where Image analysis will easily pick up many similar features (because the same blot is hybridized twice). But in this case, reprobing is an appropriate and informative analysis, rather than an indication of a problem in figure preparation or presentation.
Thank you, I really appreciate your thoughtful response. I also considered this as a possible explanation for the appearance of similar patterns in the +/+ lanes of both blot images before posting my first comment. I felt that the absence of Northern blot protocol in addition to other concerns merited bringing this detail to PubPeer for a discussion. I will outline four concerns that still remain and describe them in detail below:
Before expanding upon these four concerns, I need to describe another unexpected similarity between the two experiments. In the images below, a rectangular region of background (enclosed within a blue-dotted rectangle) is found in the upper right corner of the -/- lane of both experiments. In both experiments’ images, the background pattern that surrounds these unexpectedly similar rectangular features is completely dissimilar between the two experiments. In the exon 1-17 experiment, this region lies directly above the horizontal line and bright vertical feature that I identified in my first comment; I will address this further in Concern 4.
Returning to my remaining four concerns:
The protocol for the Northern blot experiments was not reported, so we do not know if the same membrane was used for the experiments with both probes. If the two experiments used different membranes, then the presence of so many unexpectedly similar features in the two experiments’ blots is very unexpected. If the same membrane was re-used (yes, this is an acceptable procedure), was the membrane stripped before re-hybridization? If the membrane was re-used but was not stripped, then this could explain the persistence of similar features in the +/+ lanes in both experiments, but it would raise the question of whether or not the employed protocol was capable of discriminating between signals from the two probes in the experiment where the membrane was re-probed. If the same membrane was re-used and stripped, how likely is it that such a high degree of similarity is present in the +/+ lanes, not only in the background but also in the mRNA bands?
There is an unexpectedly high level of conservation of detail in both experiments’ +/+ lanes. It is not only the presence of very similar distinguishing features in the +/+ lanes of both experiments but also the precise conservation of detail in these features that is unexpected (the findings I reported in my first comment are supported by Forensically’s clone analysis, below). Perhaps this can be explained if the same membrane was re-used in both experiments but not stripped before re-probing, and if both probes have identical hybridization efficiency? Would even a perfect and identical hybridization efficiency between both probes and Unc5b wild type transcripts generate the observed precision in conservation of detail (this also requires that no error was introduced with film placement during exposure to the probed membrane)?
There are no similarities between the +/m lanes of the two experiments. These samples contain mRNA extracted from heterozygotes bearing one wild type Unc5b allele, and the heterozygotes were capable of transcribing wild type Unc5b mRNA. Additionally, the authors reported that “all heterozygous animals were indistinguishable from wild type” and the mutation is “expected to generate a severely hypomorphic or null allele” and thus prevent wild type Unc5b transcription. From this, it is reasonable to expect that the signal reported in the +/m lanes of both experiments comes primarily from the wild type transcripts that the heterozygote can produce. As discussed in Concern 2 above, if the similarity in patterns seen in the +/+ lanes of the two experiments is due to both probes having the same high level of hybridization efficiency with the wild type transcript, then there should also be many similar features with highly conserved detail in the +/m lanes. This is expected particularly if the same membrane was used in both experiments. However, I’ve double checked these +/m lanes and I can’t find any similarities in distinguishing features.
A horizontal line, a bright region, and an unexpectedly similar background region are present in the -/- lane of the experiment for exons 1-17. The re-use of the same membrane in both experiments and wild type transcript-probe hybridization efficiency do not clearly explain the presence of these features. If this line is a membrane imperfection and the same membrane was re-used with both probes, then this feature should be present in both experiments’ images? An explanation is also required for the presence immediately above this horizontal line of a rectangular region in the exon 1-17 experiment that is unexpectedly similar to a region of the exon 5-17 experiment. These concerns are especially pertinent due to the location of these three features in the immediate region of the band whose presence – as you pointed out – is expected only in the experiment that probed for exons 1-17.
These are necessary questions due to the essential support that these Northern blot experiments provided for the work reported in this article. The protocol used by the authors could address some or all of my concerns, but the absence of a detailed report of this protocol makes it important to investigate the basis for features whose presence is not expected in the reported experiments.
I agree that if reprobing of same membrane helps explain similarity of +/+ lanes, you might expect more similarity in the +/m lanes as well (your point 3).
I also agree that the features you have highlighted in the m/m lanes are unexpected (your point 4)
I hope the authors of the study will reply to these comments.
In order to expand upon concerns 2 and 3 (described above), I highlighted more unexpected similarities in the +/+ lanes of the experiments for exons 1-17 and 5-17. I colour-coded features that are common to both +/+ lanes by using the same colour to identify each unexpectedly similar feature. Although I selected only oval regions for ease of comparison, further examination of regions outside these ovals indicates that the entire +/+ lane is unexpectedly similar in both experiments.
In both experiments’ images, regions enclosed in numbered magenta ovals highlight a linear boundary between the +/+ and +/m lanes. Each magenta oval is bisected by a yellow line. The yellow line marks the point where the unexpected similarities between the two experiments’ images end: on the left side of the yellow line is a region where unexpectedly similar features are shared by both images, and on the right is a region where there are no similarities between the two images. Note that the +/m lanes of the two experiments are completely dissimilar.
Red boxes enclose the rectangular region in the m/m lane that is unexpectedly similar in both images (described above).
Two more figures have noteworthy features:
Editorial Expression of Concern Published: 18 December 2023 Editorial Expression of Concern: The netrin receptor UNC5B mediates guidance events controlling morphogenesis of the vascular system
https://www.nature.com/articles/s41586-023-06944-2
Editorial Expression of Concern to: Nature https://doi.org/10.1038/nature03080 Published online 27 October 2004
Nature is publishing an editorial expression of concern on the article “The netrin receptor UNC5B mediates guidance events controlling morphogenesis of the vascular system”, by X. Lu. et al., to alert the readership that image integrity issues have been raised with some of the data. A report to the Stanford University board concluded that this article reflects evidence of manipulation of research data. We have considered the issues and find that the far-left lanes of the top and bottom panels of Supplementary Figure 2E appear to be duplicated. Given the age of the article there is insufficient information to conclusively assess how the duplication occurred. However, the authors have provided contemporaneous replicates that confirm the validity of the data presented in Supplementary Figure 2E.
Xiaowei Lu, Ferdinand le Noble, Frederic De Smet, Jean-Léon Thomas, Monica Autiero, Peter Carmeliet, Marc Tessier-Lavigne and Anne Eichmann agree with the publication of this statement.
Li Yuan, Quingjan Jiang, Benjamin de Lafarge, Daisuke Sugiyama, Christiane Bréant and Filip Claes have not responded to correspondence about this statement.
Attach files by dragging & dropping, selecting them, or pasting from the clipboard. Uploading your files… We don’t support that file type. with a PNG, GIF, or JPG. Yowza, that’s a big file. with a file smaller than 1MB. This file is empty. with a file that’s not empty. Something went really wrong, and we can’t process that file.
Comment must be at least 15 characters.