Monday, July 27, 2009


Whirlwind final three days at work prior to my vacay. Trying to wrap up the last of the reviewer comments for my 100+ day peer-review manuscript. Why must they suggest removing the most difficult figures/tables to replicate in the text? They're big because they condense a lot of information into a single figure/table! Like they say, a picture says a million words (or a million data points as the case may be), no? Of course, that's the worst of the review, so I should be thankful. Here is hoping this is accepted by the time I return from my lighthouse gazing! Then received a rough draft of a paper I'm doing with a collaborator in which my lab performed a fair number of RISA's. It was the first time we used this particular RISA protocol, and we're publishing results done with the protocol a couple of times this year, but I have to admit ... despite my reasons to use RISA over T-RFLP to begin with, we're switching to T-RFLP, which seems a whole lot cleaner. I'm sure my lab appreciates me having them perfect this RISA protocol over three or so months only to change directions nine months later. Besides, using a designed pair of ligated linker ends, getting sequence on these fragments is now possible. That sort of eliminates RISAs main "pro". Anyways, have to modify a couple of my figures, which is no biggie now that I've finally managed to nail down Photoshop Elements (thanks confocal microscopy class!). Good thing I saved the TIFs with their prospective documenting layers.

And I want to do the Isis transcript meme! I just need to find a copy of my undergrad transcripts. Wonder if HR will give me a copy.


Genomic Repairman said...

I don't get why they would rather have a convoluted and confusing text rather than a nice and neat figure that can summarize the data. Somedays I wonder if the reviewers took the shortbus to manuscripts some times. I one time got a comment that said, "I don't like the figure but it works." I could give a shit if he/she, their kid, or their Armenian cousin likes my figure if it works it works.

Philip H. said...

most of the time, I think such comments are really saying "That's an asweome, easy to understadn representation of the data; because I didn't think of it, you need to take it out so no one else discovers its awesomeness."

Thomas Joseph said...

GR: I don't know either, but now my manuscript has a pretty figure AND a section of convoluted text. The problem is, while we reproduced the data from prior studies (by which we compared our own) and calculated the slopes of their processes (and did the same for our own) they were only directly comparable at a few points on the graph. So our text picked those points to do direct comparisons (where we outperform past processes by anywhere from 150 to 800%), but still, the figure does the best justice to all the data so it MUST remain as far as I am concerned.

As for the suggestion to remove the table. I just axed it. If people complain that they can't follow along now (or rather, of they want to follow along completely they'll need to build their own table ... NOT MY FAULT! Of course they'll never know that that table was there to begin with, which would have made life so much easier, but I'll have to live with it.

Just because I cover 95% of the information in the text, doesn't mean that the table is not necessary. It's there because it makes following along easier if you have something to refer to, especially when half of what I talked about is found three pages back, buried in the text you probably barely skimmed over.

Why just ax the table? I fought a couple of other battles with data the reviewers didn't think should be included (guess they thought I was extremely thorough to begin with), and they happened to be things I deemed 100% necessary. So I deleted what I thought I could live without. Tried to choose my battles wisely and all that.

Philip: Heh, that is so very true most times.

soil mama said...

hmmm, I want to hear more about your thoughts on ARISA vs T-RFLP. our lab sticks with T-RFLP then does a limited clone library and in-silico digest to attempt to name dominant peaks (which hasn't worked very well for me, especially for bacteria- the insilico peaks rarely match the sample T-RFLP peaks). I would love to learn how to better name the fragments

I just finished the workshop on analyzing large community data sets and have learned some cool new tools, and one that makes PCord look EASY! once the rest of the conference is over, I'll organize my thoughts and notes, I'll send some info your way.

have a good vacation!