IMAGE ANALYSIS OF BLOTS AND CAT ASSAYS

M. L. Dodson, Vanderbilt Univ. BDODSON%VUCTRVAX at PUCC.PRINCETON.EDU
Wed Dec 12 09:35:00 EST 1990


andy mendelsohn asked about using a scanner to quantitate blots.  This is our
experience on that subject.

We use a Microtek gray scale scanner and scan analysis.  If you are willing
to fiddle around enough, you can get some quite nice results, BUT be prepared
to fight the software all the way.  The host of this stuff is a Mac II.  Firstly
   ,
DON'T overexpose the blot (of course the same goes for using a densitometer,
as a matter of fact, even fighting the software, this beats using a
densitometer).

When you use the grayscale scanner, cut the contrast down as low as possible
to increase the dynamic range (same reason as avoiding overexposure).

We save the result as a TIFF file, then go into Scan Analysis to integrate
the image.  You can tie all these disparate programs together, maybe with
Hypercard; we use ProIcon, a string and list processing language which is
also useful for sequence analysis and for writing filters to take text output
from our VAX and prepare it for, e.g., Excel.  Check it out at anonymous
ftp from cs.arizona.edu; cd to icon; download the readme files.  But, back to
scan analysis.  The program (available from BIOSOFT in the UK) suffers from
a very misfortunate user interface: hopelessly rigid in all its automatic modes.
We just use the manual mode which allows you to draw boxes around the areas you
want to integrate, then replicate the same size box elsewhere on the page to
do other blots.  The program just sums the grayscale results for each pixel.

The hooker is that grayscale images are inverse to what scientists are used
to:  black has a lower number than white.  You can deal with this in Scan
Analysis, but we just handle it in Excel.  You can paste the results from
Scan Analysis into Excel, or save it in a file, then open it in Excel as a
delimited file.  We find it useful, since we are using ProIcon, to filter the
Scan Analysis results into a tab delimited text file, which is quite palatable
to Excel.  You can then write macros in Excel to complete the analysis.  These
macros can have dialog boxes, etc.

Sorry for the rambling nature of this reply.  I just got to work and haven't
had any coffee yet.

Scan Analysis:  BIOSOFT
                22 Hills Road, Cambridge
                CB2, 1JP, UK

or (I just noticed they have a US distributor)
                P.O. Box 580, Milltown
                New Jersey, 08850

Hope this helps.
Bud Dodson
Center in Molec. Toxicology
Vanderbilt




More information about the Bio-soft mailing list