Reader Comments

Post a new comment on this article

Still confused

Posted by jkhartshorne on 17 Jan 2014 at 05:05 GMT

You say that your contribution is "to make these types of methodologies more accessible for researchers from different scientific backgrounds and without the need for strong computing skills," and you provide R code. But then you cite the R package "fields", which has been around since 2006 and which already has a quilt plot function.

No competing interests declared.

RE: Still confused

hwand1 replied to jkhartshorne on 17 Jan 2014 at 06:15 GMT

Dear Reader,

Thank you for your comments. If you notice that R package- fields was just compared "broadly" with the "quilt plots"; that's why it was cited. Regards

No competing interests declared.

RE: RE: Still confused

noamross replied to hwand1 on 17 Jan 2014 at 18:46 GMT

What is the "broad comparison" when the contribution is nearly identical to quilt.plot? (Which was introduced in Feb 2005). One can recreate the plot in the paper with just a these lines of code:

library(fields) #load the fields package
x = rep(1:10,3); y = rep(1:3, each=10); z = rnorm(30) #creating some fake data
quilt.plot(x,y,z, nx=10, ny=3, col=two.colors(start="yellow", end="red", middle="orange"))

The author's code could be modestly helpful, in providing labeling functions. But any gain is probably offset by the fact that they do not even provide the code in a usable format. Instead, it's provided as unusable *screenshots* in the supplemental information.

No competing interests declared.

RE: Code for method is non-reusable

hlapp0 replied to jkhartshorne on 20 Jan 2014 at 03:35 GMT

I think this comment in the first comment in this thread bears repeating:

You say that your contribution is "to make these types of methodologies more accessible for researchers from different scientific backgrounds and without the need for strong computing skills," and you provide R code.

Except that you don't provide R code. You provide screenshots of R code, which makes your method not more, but much less accessible, in fact inaccessible to users. I ask the authors of this article how they reconcile this with the claims in the text, and I ask the editor how this wasn't caught in peer review. I also ask the editor whether and how any of the code, and thus the reported claims in the paper were actually verified? Did the reviewers have access to the code in text form rather than in the form of screenshots?

Is this not a case of the provided evidence not supporting the claims being made? And given the evidence we have, how can one believe that any of the claims have been verified? And thus, is this not a case of unsound science?

No competing interests declared.

RE: RE: Code for method is non-reusable

tthatcher replied to hlapp0 on 21 Jan 2014 at 16:05 GMT

Novelty is sometimes a difficult concept to evaluate. If this is a subset or special case of an existing method, presented as a useful method for researchers in resource-poor environments, it might have value even if it is not novel. However, under PLOS ONE's commitment to open access, as well as its editorial policies, the actual R code should be made available in a public repository so that researchers in resource-poor environments can access it freely. Will the authors consider depositing the code at PLOS ONE as supplemental data or making it available at a free public repository?

No competing interests declared.

RE: RE: RE: Code for method is non-reusable

neilfws replied to tthatcher on 21 Jan 2014 at 21:16 GMT

I do not understand what is meant by a "useful method for researchers in resource-poor environments."

The code in this article requires that users can install and use R. As do the numerous, existing methods for generating heat maps using R. I fail to see how this code is more or less useful for R users than any other R code.

No competing interests declared.

RE: Code for method is non-reusable

tthatcher replied to neilfws on 21 Jan 2014 at 22:16 GMT

With regard to the actual value of the manuscript, I don't personally know how to use R, although I do know who to go to to ask for help if I needed it. I can imagine that there would be value in publishing, for example, a set of R macros or scripts that a person with low access to R experts could use to derive useful graphical outputs of complex data sets. (Remember that P1 may be accessible to researchers in parts of the world that may lack the supporting infrastructure available in the US or Australia.) To me, that would be a blog post, not a published P1 paper. But I wasn't the handling editor.

On the issue of retraction, it has been noted elsewhere that retraction is appropriate when an article is factually wrong, or the result of misconduct or other ethical violations. That's not alleged here, just that this is not novel enough to have been published in the first place. If the issue here really is just, this paper falls below the "minimum publishable unit" then that's a judgement call, and maybe a reason to reevaluate standards and processes going forward, but not a reason to punish the authors with a retraction of a good faith manuscript.

No competing interests declared.