SmartFlare Paper Revisions

Not too long after publishing version 1 of the SmartFlare paper at Science Open, we received a couple of very helpful reviews. We’ve been working on the points made by the reviewers to improve the manuscript for version 2 (coming very shortly!). Below are some thoughts on a couple of the comments and a quick update on some new data in the works.

You can find the reviews here.

VEGF mRNA levels +/- DMOG

One point that came up in both reviews, was that we had not explicitly shown that treatment with dimethyloxalylglycine (DMOG) increases VEGF mRNA levels. We originally referenced a couple of papers where it has been used, although these (it was rightly pointed out) were in cell-types other than the HeLa cells we were using in our study.

Requesting this control is a fair and important point. We do not see any change in SmartFlare fluorescence in cells with or without DMOG treatment (500nM for 18h alongside the SmartFlare addition), so it’s important to know that VEGF mRNA is actually changing under those circumstances.

quantitative PCR is as the name suggests is a way to measure your mRNA of interest, which is reported relative to a stable unchanging “housekeeping” mRNA (similar to a loading control in Western blots).

The Sée Lab has a lot of experience with DMOG, VEGF mRNA and qPCR so we asked Sophie Cowman to help us out. In three independent replicates we found that in in our cell system, treating with DMOG (500nM for 18h), leads to 20.63, 15.26 and 21.68 times (mean: 19.19) more VEGF mRNA (compared to untreated cells).

You can find more details of the protocol including the Cyclophiln A housekeeping gene in this paper, or wait for version 2 with all the details.

Presenting ‘Typical’ Images

One of the real differences in our SmartFlare paper (compared to almost every other paper in which they’re used) is that we’re trying to study from where in the cells the signal is coming. This, as with any study dealing with lots of imaging data, leads to the problems of how to present those data.

Let’s take as an example one of the first experiments we did with the Uptake Control, finding that the uptake was heterogeneous. If we look at a whole frame (about 150 microns square).


In most publications, panels like this would get shrunk down until you couldn’t make out the puncta, so you get an idea of the overall intensity level but not the distribution. The alternative is to look at the distribution and lose the big picture. If we take a ‘representative’ cell from the above panel we get something like this:


While we are representing the ‘typical’ cell (at least by majority), we don’t see what we’re trying to demonstrate. I’m not sure that there’s a good solution to this problem. Montages of many fields are a possibility but then this only really works for a digital medium where you can zoom in and inspect the data. Our way of solving this problem is to provide exemplar images in the paper and to link out to our Open Data repository where people can inspect, zoom, contrast enhance and analyse to their heart’s content.

Response to Reviewers

We’re almost finished the manuscript changes, at which point, we’ll push the updated version and add a full response to the reviews on the Science Open page.


3 thoughts on “SmartFlare Paper Revisions

    • Fair point, but it raises the question of how many figures are significant for ratios of qPCR data. The value of 19.19 fold increase could probably be described as 19 fold and still be defended as being precise enough. What about 20 fold?

      Answers on a postcard…


  1. Thanks for the update.
    One point that I raised in my review is that you could be already in saturation, so an increase by 2-fold or 20-fold – it will not make a difference for SmartFlare. Assuming that there is saturation it still does not bode well for SmartFlare as a good tool to measure expression levels since it will mean it has a low dynamic range.
    Or it could be that SmartFlare just does not recognize the RNA as you suggest.

    regarding the other issue – due to limited space in typical journals you should show a typical image alongside quantitative data from you entire image collection. the image should be representative of the data (e.g. if in most cells you see co-localization of x & y, show an image of such a cell).
    putting the images as open repository is nice, but will not help the casual reader. not everyone has the time or the ability to analyze your images.

    Can’t wait to see the revised version.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s