 |
 |
 |
 |
 |
I've used the routine of copying and pasting dot plots and contour plots
into graphics programs like Photoshop, but I've developed a different
technique for getting histograms into publishable quality.
There's a shareware program from Sumex-AIM (a share/freeware MAC site)
called DataThief. The latest version is 2.0, although I seem to have
better luck with version 1.0.8. The purpose of this program is to take a
PICT file of a graph and convert the line graph into X,Y coordinates that
are then saved into a tab-delimited file format. Once you have each of
your single color histograms in this format, it's easy enough to use Excel
or CricketGraphIII to make nice, publication quality plots. Your plots can
be overlayed, as in the case of negative and positive populations, and now
that your plot's in a graphing program, you can do anything you want with
the axis labels, titles, legends, etc.
Some Tips:
1) Make your plot in CellQuest as large as possible. I think histograms
come out better with a little smoothing (about 5). The larger it is, the
more points there will be to export (this is also true when copying dot and
contour plots into your graphics programs). Generally I make the document
size 2 pages wide and use the whole area for the plot. I use the Cell
Quest zoom feature to make the actual histogram take up as much of the plot
as possible.
2) Copy the plot and paste into PhotoShop. Where the graph line
intersects the axis, you will probably have to erase the x-axis on the
right side. The auto-trace feature in Datathief will follow the lowest
path, so if your line hits the axis, your data from there on out will be
just that, zero...
3) Select AutoTrace from the righthand DT menu. In file options deselect
"Include Headers." (if its on it puts some notations on the top of the
exported data that you don't really need.) In Trace options, make the
trace height around 50. If you're histogram has very steep and long
portions, you might have to increase this number. Select for a logarithmic
X-axis after you've entered X-min,X-max, Y-min,Y-max.
4) When making your graph in your graphing program, choose the x-axis to
be logarithmic, or else it will look strange.
If you play around with all of this, you should come out with some
excellent looking graphs. The one last difficult step is if you want one
histogram, say a negative sample, to be a grey shade, and the other one
shaded white. There must be a graphing program that does this, perhaps
DeltaGraph or SigmaPlot, but I'm so familiar with CricketGraphIII, that I
use that program preferentially. If I want shading, I copy the plot that
I've made in CGIII and paste it into Canvas. The ungrouped plot in canvas
consists of 1)blocks of text, 2) the axes and tick marks, and 3) the plot
line - as a series of short, connected lines (essentially a multifacetted
polygon.) To shade an area in Canvas, that area has to be completely
enclosed by a line, so you have to edit one of the end points and connect
it with the other (since the axis is exported separately.) A little time
consuming, but the final product comes out very nice.
I hope this helps. I wish it didn't consume so much time but when we used
the HP with Lysis and the paintjet printer, I used to scan in the printed
histogram, so at least I've eliminated a step.
Todd A. Johnson
UC San Diego, School of Medicine
tajohnson@ucsd.edu
>I have used the same technique that Derek describes, of copying the plot to
>the clipboard and pasting into a drawing program. I totally agree that BD
>needs to add an export function to CQ, one that allows export as PICT,
>TIFF, or other formats.
>
>When the plot is pasted into another program, it is a bit-mapped image,
>with several layers grouped. One consequence of bit-mapping is all text
>prints poorly, like using a dot-matrix printer. For publications and
>slides, I have edited those out and added my own labels to the axes.
>
>To do that, one must first ungroup the figure (I use programs like Canvas
>and ClarisDraw, but others are likely to work, as well). I was immediately
>surprised to learn that the axes and labels are in one layer and the data
>(dots or contours) are in several other layers, apparently corresponding to
>"cuts" in the data. So, for a contour plot, there is one layer that
>includes all contours, another that includes all but the inner contours,
>another that includes all but the inner 2 contours, ad nauseum. One way to
>speed up printing, is to send the top layer, that includes all data
>(contours or dots), to the back, then delete all other layers. (As I
>recall, the axes and labels are the top layer, so that must also be sent to
>the back, even to be able to select the various layers of dots or contours.)
>
>I echo Derek's recommendation of "getting the plots into as near the
>finished article as possible before exporting." Because the pasted figure
>is bit-mapped, any resizing will increase the size of dots or thickness of
>lines. It will look ugly. When I want to group several histograms for
>export, I use a single plot in CQ, that I have adjusted to the final size
>that I want. Then I load the data for each sample into the same plot, to
>assure that all plots will be the same size in the finished version.
>
>Also, I would note a common error the I make is to select the plot
>incorrectly when copying. If you click on the plot, not on the border
>around the plot, the copy command is not active. I have been surprised at
>times, when I thought I copied a plot, then pasted something altogether
>different into my drawing program.
>
>Hopefully, BD will implement a better way in a future release.
>
>Roger
>
>>Hello there,
>>
>>I have had a lot of trouble with this very problem. It does seem
>>ridiculous that there isnt the facility for exporting CQ documents as a
>>graphical format.
>>
>>Using the clipboard is fine as long as the program you want is on the same
>>computer and you have time to play with it etc. The approach I am using is
>>to use an intermediary program to dump the CQ graphics in and then save as
>>a PICT file.
>>
>>I am using a shareware program called Graphic Converter which is
>>obtainable from the Info-Mac archives at Sumex. The location (off the top
>>of my head) is:
>>
>>gst/grf/graphic-converter-25.hqx
>>
>>This enables one to paste the dot-plots, histograms or WHY into a document
>>which can then be saved in any number of commonly used graphical formats
>>eg BMP, GIF, JPG, PICT and TIFF. As far as Im aware, most common programs
>>used to generate slides such as Photoshop, PowerPoint, Claris and Word can
>>read PICT files so that is what I normally recommend (I believe filesize
>>is smaller as well).
>>
>>Even so, I would still recommend getting the plots into as near the
>>finished article as possible before exporting.
>>
>>Hope this is of some help, but if anyone is using a different approach I'm
>>sure we'd all be happy to hear about it.
>>
>>Derek
|
|
|
|
|
|
CD-ROM Vol 3 was produced by Monica M. Shively and other staff at the
Purdue University Cytometry Laboratories and distributed free of charge
as an educational service to the cytometry community.
If you have any comments please direct them to
Dr. J. Paul Robinson, Professor & Director,
PUCL, Purdue University, West Lafayette, IN 47907.
Phone: (765)-494-0757;
FAX(765) 494-0517;
Web