Reader Comments

Post a new comment on this article

Data quantification concerns

Posted by PLOS_ONE_Group on 19 Jun 2013 at 23:41 GMT

PLOS ONE staff editors removed a comment on this article due to it not meeting our commenting guidelines (available at http://www.plosone.org/st...). An edited version of the comment is copied below.

Matt Hodgkinson
Associate Editor, PLOS ONE

--

Data quantification concerns
Posted by jamesmcd on 08 Jun 2013 at 03:13 GMT
Plos One disappointed me. This paper has some serious issues with its images […].
Given that the authors have provide one original blot on this correction, I've quantified them. For instance, in the new representative figure 6, the OD for each band are:
C 39191.522
C 35009.794
CD 35190.744
CD 40882.401
D 34375.622
D 27429.501
DS 31716.886
DS 16388.329
The averages of those OD do not even closely match the graph 6A. Worse yet, they lead to the opposite conclusion. Therefore the figure is not representative, or the graph is wrong. The correct graph for that western blot can be found at http://img580.imageshack..... Please notice I kept the spelling error from the original graph, that survived the peer review.That said, the conclusion is biased, and cannot be trusted. In this case, as it has been accepted for publication, the very peer review process from Plos One appears to be questionable

Competing interests declared: PLOS ONE Staff
report a concern respond to this posting

Comment posted on behalf of the authors

PLOS_ONE_Group replied to PLOS_ONE_Group on 28 Aug 2013 at 15:47 GMT

Dear readers of PLOS ONE,

I would like to clarify how the values for Figure 6A were obtained. The densitometry analysis was performed following the instructions for the ImageJ Program (rsbweb.nih.gov/ij/docs/guide/user-guide.pdf). Briefly, the following steps were performed:

1) Images were scanned, saved as TIFF and gray scale for analysis;

2) Images were opened using the ImageJ Program;

3) The first lane was outlined using the "Rectangular selection" tool;

4) The first lane was selected by click on "Analyze → Gels → Select First Lane";

5) The rectangular selection was moved over to the adjacent lane; the outlined lane was then selected by clicking on "Analyze → Gels → Select Next Lane";

6) The previous step was repeated for each remaining lane;

7) The lane profile plots were generated by clicking on "Analyze → Gels → Plot Lanes";

8) The areas under the peaks were marked off by using the "Straight line" tool;

9) The areas under the peaks were determined by clicking inside them using the "Wand (tracing) tool”;

10) All procedures previously described (1 to 9) were performed for: a) band densitometry corresponding to the protein of interest after Western blot analysis and b) relative amount of protein in the band using the Ponceau red method, from six animals per group;

11) Values for band densitometry of the protein of interest (a) were normalized (divided) by the respective relative protein amount (b), as described in the manuscript. Normal Gaussian distribution was assumed and outlier values detected using the following criteria: mean (μ) = 0; variance (σ) = 1; tail area (a) = 0.05; and critical value (z) = NormsInv (0.95).

This is the reason why the values presented in the "Reader's Comments" (from two animals per group and with no normalization) do not properly reflect the final result of the quantitative graph that has been published in the manuscript.

The full data for the results reported in Figure 6A are available via the files below:

http://www.plosone.org/at...
http://www.plosone.org/at...

Competing interests declared: PLOS ONE Staff
report a concern respond to this posting