Friday, June 16, 2006

Making Figures Sucks, or I Suck At Making Figures

Ugh. Maybe I just have an insecurity complex, but my thesis advisor always told me I sucked at making figures, and I think he was right.

Even if I didn't suck then, I think his saying that has made me want to quit trying. Now I look at my figures and I hate the layout and the labels all look bad, but I feel like I have no talent for this stuff, that it's the sort of thing I never learned to do in kindergarten. No matter what I do it always comes out looking sloppy.

For example, the gels are never perfectly straight, even when I align them using guides in Photoshop, everybody always seems to think it could go just slightly more one way or the other. This would be fine, but there's no right answer, since the lanes on the gels didn't run perfectly straight to begin with.

Then, if I have any pictures of cells, the contrast never comes out right, but I'm afraid to play with the contrast too much for fear of someone accusing me of altering the data. So that's a lose-lose situation.

My advisor was great at this stuff, mostly because he was obsessive and willing to spend hours lining things up to the exact pixel. But I think he didn't do me any favors by cleaning up my figures in grad school, since I still have no patience for it, and I have to wonder if there aren't some expert tricks I could be using to make my stuff look better?

At one point I tried to read some books on graphic design and layout, but I never found them very helpful. I still don't know how to get around the problems of formatting figures for column widths AND presenting things in a logical order AND having them be the right size, nevermind having the whole figure look pretty at the end when there are 6-8 pieces to each one. Yuck! It always ends up looking crowded, the labels are too small or the fonts don't match and I have to go back and re-label everything, especially when I keep changing my mind about what pieces go in which figures.

Why isn't there a whole industry devoted to this, if there are hundreds of scientific journals out there publishing zillions of figures every day? I really want somebody to invent a magic molecular weight labeler. When I draw the lines on the film, they're never straight. This doesn't bother me until I scan in the gel and have to decide, is it worth drawing them in, one at a time, with a nice perfect rectangle? This is not what I want to spend my time doing.

13 Comments:

At 7:27 PM, Anonymous Anonymous said...

You neeed to get out more.

 
At 7:47 PM, Anonymous Anonymous said...

Please don't. I realize this goes against get-the-paper-into-a-10+impact-journal-theory, but a gel is a gel. I can tell what's what from any autoradiogram or fluorescence image, and any undergrad can do the same. I would much rather see your artifacts at high resolution than your artifice at low resolution. Sometimes I land on the extreme side of this and refuse to run short oligo (10 or less) gels with a ladder, because if you can't tell n from n-1, there is no need to confuse yourself. Clean gels make me inordinately suspicious. When I bring mine to group meeting I deliberately leave residual 32P-ATP and the lane loading residue visible so everyone can know what I'm up to. Better to let everyone see what I've got going than give anyone unrealistic expectations. Oftentimes I will overcontrast printouts of imaging of my P32 gels so people can see the smear if my loading was sloppy, because, hey, it affected the integral. Warts and all, baby. Anyone who tells you to do otherwise isn't worth it. Science doesn't happen in rectangles, it happens in blobs.

 
At 9:06 PM, Anonymous Anonymous said...

A non tenure instuctor person in our department who is in her early 30's is really into being called "Dr." I recently exchanged several emails with her and each time they were signed Dr, even though I signed my first name. And she is actually writing Dr., it's not like the email program is automatically doing it. She is Austrailian, but I didn't think that they were really formal people. I asked one of the grad students about this and she said that she noticed it too. She seems like a nice person, but when I see her in the hallways should I call her Dr. and her last name? It's kinda weird.

 
At 2:45 AM, Anonymous dekay said...

In general what you may want to look at:

http://www.edwardtufte.com/
http://groups.google.com/group/comp.text.tex/browse_frm/thread/bf32d29202c23a11/53f335c517aede5c?q=thesis&rnum=1#53f335c517aede5c
http://presentationzen.blogs.com/
http://www.infosthetics.com/

Just focus on the point you want to make (and don't include more than that in the graph).

 
At 10:13 AM, Blogger Ms.PhD said...

Yes, I do. Unfortunately I can't get out of being a postdoc unless I publish a couple more papers.

re: making gels look ugly just so people will believe you... I used to be one of these people. Unfortunately my gels would take up pages and pages of space if I showed the whole thing, and there's nothing to show- I finally learned how to run a relatively clean western (she says, hoping not to jinx the ones she has to develop today!). But yes, I am with you on the blobs. Nicely put.

Let her call herself Dr- it's probably a phase and she'll get over it when she realizes nobody else does it. Most of us went through a short phase of self-glorification when we first graduated, because really, what do you have to show for yourself besides the letters of your degree? Call her 'hey you'- that's what I do with people when I can't actually figure out what to call them. Start your emails with just "hi". Email etiquette leaves lots of room for flexible professionalism.

Dekay, I read Edward Tufte's book(s) a few times over the years and mostly I find them too abstract for what I want to know. I can easily see why one way of presenting one type of data is better than another. My issues are mostly with layout, and not the LaTeX sort (and it's not a thesis I'm writing, though that one link would have been useful about 5 years ago when I was). I'm going to have to read the presentation blog more, although I tend to do better with talks than with papers. And the infosthetics thing looks marvelously cool! I love the handshake bracelet info-exchanger! Totally Jetsonian.

Re: focusing on the point, I think that's the problem. I put off writing this paper for so long, now I've accumulated a lot of information that should go in but I'm not always sure where to put it. I always have this problem of not knowing when to stop until I've gone a little past where I should have exited the highway.

Emphasizing the major points while not excluding the minor ones (or worse, overemphasizing the minor ones) is quite a balancing act, and I don't think I'm very good at it.

Writing-wise, I can eventually get there if I go through enough drafts, because I like revising writing, always have. Drafting and revising figures, on the other hand, is way up there with getting my toenails pulled out. I guess I should just view it the way I do writing, and keep ripping them apart and reorganizing, but it's really laborious.

I think I want somebody to give me a cookbook of rules to help me set up the layout from the beginning so it doesn't come out looking kludged. And I don't want to have to wade through Edward Tufte's books to get the recipes.

 
At 1:03 PM, Anonymous Anonymous said...

Sound like anyone (or everyone) you know ?

"Egghead: A person of spurious intellectual pretensions, often a professor or the protégé of a professor. Fundamentally superficial. Over-emotional and feminine in reactions to any problem. Supercilious and surfeited with conceit and contempt for the experience of more sound and able men. Essentially confused in thought and immersed in mixture of sentimentality and violent evangelism. A doctrinaire supporter of Middle-European socialism as opposed to Greco-French-American ideas of democracy and liberalism. Subject to the old-fashioned philosophical morality of Nietzsche which frequently leads him into jail or disgrace. A self-conscious prig, so given to examining all sides of a question that he becomes thoroughly addled while remaining always in the same spot. An anemic bleeding heart.

 
At 1:05 PM, Blogger alamode said...

I'd suggest to use Photoshop to make all your final figures. Use one layer for the background graphics and another for the text, arrows etc. Always save as a .tiff with the layers so you can go back and edit and rearrange things fairly easily. When you are done, flatten everything and save as a "final version" but keep the layers version for editing.

Those instructions from the journals are crazy sometimes, but basically be sure to save each figure as a 3.5 or 7 inch wide by 9 inch (max) tall figure at 300dpi. They can be large files, thats life.

Adjusting brightness and contrast are no trouble for most journals and they say so. Scaling and cropping are expected. Just dont go erasing "bad" marks or selectivley enhancing a single band on a
western and you will be fine. (Plus, no one wants to see your blots anyway, check out any recent JBC, Science, Cell, etc.. and look at the teeny tiny westen blots that only show the band of interest).

As for creating a nice looking graphic, just look at the journals you like, and the figures you admire, and try to emulate them. No shame in that.

Just my 2 cents.

 
At 7:15 AM, Anonymous Anonymous said...

I love making figures. How about you clean up my writting and I'll do your figures??

 
At 4:49 PM, Anonymous Anonymous said...

Ahh...doing a postdoc

Work for 2 years at low pay. Your first Western looks great, then its 3 weeks of 1. pos cotrol doesn't work
2. negative ctrl is positive
3. reagent runs out

Then...you have to stop all work, and write a grant proposal. Then, you have to stop everything again b/c you have to give the department seminar.

Fast forward 2 years. You submit your work, it is read by 3 indifferent or nasty reviewers, and sent back for revisions. Once its is finally accepted, a few people read it, and maybe the 1 or 2 labs working directly
on the same pathway or protein bothers to repeat the experiments, or modify their plan based on your results. The other 99% of journal readers wither do not notice your paper, or read it once-- possibly just Fig 7 and the discussion-- and then flip the page. Or turn on the TV and completely forget ehat you have so painstakingly discovered.

Yay!!

 
At 1:04 PM, Anonymous Anonymous said...

The Dr. thing might be because she is Australian. I am from Canada, and everyone there calls each other Dr. I was a bit surprised when I moved to the states and saw how informal everyone is, and I'm a pretty informal person. Now when I go back to Canada, or visit other countries,I really notice how formal they are. Reminds me of that skit 'Dr., Dr., Dr., Dr." where they are all nodding Dr. to each other.

Dr.

 
At 2:41 PM, Blogger Ms.PhD said...

I do use Photoshop the way you suggest. That's not the problem.

To the person who suggested a trade- sounds great to me!

To the person who sent the Egghead quote, I'm offended- and suspect most of my readers would be?- by the use of the word 'feminine'. And I'm amused by the irony of such a pretentiously worded piece describing someone who is supposedly pretentious to a point of being unbearable.

 
At 12:52 AM, Anonymous Anonymous said...

About the egghead quote: It looks like it was written c.a. 50 years ago, which explains the language. so please don't be offended. I posted it because I thought it had an amusing historical charm to it and because it seems to still be true of many academics today, especially the ones who deny that there is a world outside university. cheers.

 
At 3:41 PM, Anonymous Anonymous said...

This probably isn't the best place to ask this- mostly because i don't want to take up the space in this blog to ask annoying questions about photoshop- but the photoshop forum i joined is not responding and i'm desparate. Here is my problem: I open up a tiff file of a field of cells in photoshop. I need to chose a cell, copy and paste it into my figure in photoshop. Then, no matter how bright the cell is in the original image, in the new image (my figure or even just a new file containing the one cell by itself)it is dimmer. Its driving me nuts. Does anyone know how i can maintain the original brightness? Thanks.

 

Post a Comment

Links to this post:

Create a Link

<< Home