Hello all,
Here's some more info regarding the material that I've asked you to
contribute to the Geuvadis main paper.
The deadline for this was today but I'll extend it
to next week's Friday, October 26. Thanks to those couple of
people who actually made it on time. If you have some super-exciting
additional analysis going on beyond that date, send me the rest and
write a placeholder of what's missing.
Text: As I said before, everyone who wants to have their
analysis included in the main paper has to send me text that
describes your methods, results and conclusions. We will likely end
up having a supplementary text that is more than just methods, and
I'll incorporate your text there, and we'll of course use this
information when writing the main text.
- Think of this as a mini manuscript, so please write proper text
with full sentences. The text can and should be detailed, but don't
overdo it; include only relevant and interesting analyses/results.
- Refer to the figures/tables you provide with e.g. Fig.
TUULI_1_samplecorrelation (your name, figure number, description) so
that we can keep track of them. Write legends as well.
- Include references to the literature as PubMed IDs
Figures: The following has been written with plotting in R in
mind - you can of course use other software, but try to follow these
guidelines as much as possible. Nice figures are extremely important
in making the paper attractive to editors, reviewers and readers.
- 1-6 figures/tables per person is probably what we can include.
- size: height=5, width=5 inches. Make sure that the plots
look good when shrinked; they will be printed in max one column
width which is much much less than a powerpoint slide or your
screen. If you deviate from the square shape, keep in mind that most
plots will be panels of a single figure and should fit together.
- background must be white or transparent!
- axis labels, legends and all other text included in the plot must
be big enough. No exceptions. The default settings are hardly
ever enough; in R, cex.axis=1.5, cex.lab=1.5, cex.names=1.5 is good
for 5x5 inch images.
- use the population and laboratory colors that have been defined in
the sample information sheet
- line width needs to be thick, 2 in most cases (lwd=2), e.g. in
boxplots, most scatterplots.
- save your figures as pdf vector graphic, in R the command is pdf()
, with the file name like TUULI_1_samplecorrelation.pdf .
- have your scripts organized so that you can reproduce figures
easily if/when needed
Please let me know if you have any questions. These instructions
will hopefully save time and nerves from all of us, so that we can
spend more time thinking and discussing about science rather than
figure formats.
cheers,
Tuuli (also known as the Figure Nazi...)
--
Tuuli Lappalainen, PhD
Department of Genetic Medicine and Development
University of Geneva Medical School
CMU / Rue Michel-Servet 1
1211 Geneva 4
Switzerland
Tel. +41-(0)22-3795550
tuuli.lappalainen@unige.ch