Densitometry of gels and immunoblots

[mhollowa at epo.som.sunysb.edu]

In article <1994Feb28.195501.1222 at galileo.cc.rochester.edu> ajp2o at crocus.medicine.rochester.edu (Anthony Persechini) writes:
>I am wondering if it is possible to do this with a flat bed
>scanner as a digitizing device together some sort of commercial or freely
>available software for the analysis. I would very much like to hear
>cost-effective methods for doing this that are currently employed by
>others. 

I have personal experience with QuantiScan from Biosoft using files made 
from an HP scanner on a PC.  It's a nightmare.  They made a graphic 
interface that is quirky, nonintuitive, and has no editing capability in 
peak marking.  This means that if you make a mistake while manually 
marking many bands, you have to start over.  The autopeak function is 
useless, even on clean blots.  Again, there is no way to edit the peaks 
automatically selected.  Instructions for selecting parameters for peak 
selection are nonexistant.  A call to tech support reveals that the 
numbers selected for peak height, etc, are completely arbitrary and 
relate to nothing at all.  There is no way of knowing what numbers have 
to be altered for the desired results.  It seems to have been something 
the programmers came up with after a late night drinking binge.  The only 
advice: play around with it and see if anything works.  Manual selection, 
for reasons known only to the programmers, compacts the peak graph to one 
side of the screen into an area 1/4 the size of every other graph 
displayed.  You then have to select beginning, middle and end (being very 
careful).  

As far as quantitation is concerned, I've compared the results to those 
of a laser scanner and have to conclude that QuantiScan will only detect 
a difference if that difference is large and obvious.  This, at least, is 
probably a limitation of the media used.