Reader Comments

Post a new comment on this article

Heat maps?

Posted by jkhartshorne on 17 Jan 2014 at 04:55 GMT

Any response to the people who say this is just a heat map?

No competing interests declared.

RE: Heat maps?

QCaudron replied to jkhartshorne on 17 Jan 2014 at 04:59 GMT

I was just going to say... I don't understand how the first author, who is trained in mathematics and statistics, has never heard of a heat map. This is not a novel concept...

No competing interests declared.

RE: RE: Heat maps?

hwand1 replied to QCaudron on 17 Jan 2014 at 06:31 GMT

Dear Reader,

Thank you for your comment. However, the authors never claimed that they have never heard of a heat map. If you read the article fully (including the discussion section) you will see that Heat map was broadly compared and cited in our paper.

Please read the Discussion section (second paragraph):

"Quilt plots" can be considered as a simple formulation of "heat maps". They produce a similar graphical display to "heat maps" when the "clustering" and "dendrogram" options are turned off.

The authors also do not claim that the methodology is new.

Please also read the last paragraph in discussion:

"Although our method cannot be considered "new", the novelty is to make these types of methodologies more accessible....."

Handan Wand

No competing interests declared.

RE: RE: RE: Heat maps?

stsmaf replied to hwand1 on 17 Jan 2014 at 10:11 GMT

For most people, I think it's fair to say, the proposed "Quilt Plot" is an example of a heat map.

c.f. where there are examples that do not involve dendograms etc.

In either case, whether the columns and rows are ordered or not, it is my impression that such plots are widely used throughout the quantitative sciences.

With regard to R code, "Instant Heat Maps in R: How-To" by Sebastian Raschka (top of Google search for "heat map" in Books) describes in some detail how to construct heat maps ("Quilt plots") in R.

No competing interests declared.

RE: RE: RE: Heat maps?

QCaudron replied to hwand1 on 17 Jan 2014 at 14:23 GMT

Fully understood. I wish I could have edited my comment, as it wasn't phrased clearly. What I was trying to ask was, "what am I missing here that made this worthy of publication in PLoS ONE ?"

If it's simply a heat map with clustering and dendrograms off - this is, I believe, an R issue. The heat map is the central part of the plot ( image(x) ), and in many other scientific computing packages ( Matlab, Python, ... ), that's all that will appear also.

Thank you for your quick response.

No competing interests declared.

RE: RE: RE: Heat maps?

eagereyes replied to hwand1 on 17 Jan 2014 at 16:27 GMT

A heatmap is simply what you call a Quilt Plot: an array of colors that represent data, typically in a cartesian layout (but can also be hexagonal, etc.). Additional data can be overlaid or added, like dendograms, and the data can be clustered, sorted, etc. But the visualization is still a heatmap.

What you've done here is mistake the heatmap plus its extensions for the underlying idea, and then removed the extensions. But that doesn't mean there's anything new here. People have been doing this for decades in Excel, for example.

I don't see your discussion making this in any way worthy of a publication, even if you don't claim novelty. This has been accessible in all sorts of software and is easy enough to code up in any language.

No competing interests declared.

RE: RE: RE: Heat maps?

r_barton replied to hwand1 on 18 Jan 2014 at 11:01 GMT

This short paper was thought to offer utility to resource-poor users, and offered a simple application example. Both the academic editor (myself) and a reviewer raised concerns about the level of originality, and this issue was specifically referred back to the PLoS ONE Editorial office, where the decision to proceed was made.

No competing interests declared.

RE: RE: RE: RE: Heat maps?

MattJHodgkinson replied to r_barton on 20 Jan 2014 at 13:31 GMT

PLOS ONE editorial staff were consulted by Prof Barton during the review process for this submission. The advice given by the editorial staff was intended to provide guidance to Prof Barton on whether his editorial handling and decision were in line with the journal’s policies and criteria and we referred to our specific criteria for manuscripts that report Methods, Software, Tools and Databases ( as a reference for the evaluation of this work. We apologize for any misunderstanding that may have arisen in relation to the advice that Prof Barton received in relation to this manuscript.

Competing interests declared: I am a PLOS ONE staff Associate Editor

RE: RE: RE: RE: Heat maps?

songqiang replied to r_barton on 23 Jan 2014 at 01:00 GMT

That is amusing. Only one reviewer questioned questioned the paper's novelty. How many reviwers are consulted?

As many people commented, there is nothing new about the so called "quilt plot". In regards for the claimed utility, the code highly depends on a specific platform R, which already requires users to be able to install and invoke R.

Finally, in case the authors, editors and other reviewers are not aware of: to make a "quilt plot", you just need to use the built-in heatmap function, and set Rowv and Colv to NA.

No competing interests declared.