jaudall at students.wisc.edu
Thu Jan 14 15:31:23 EST 1999
I'm looking for some software which would do analysis of
RFLP image files in a batch mode. Most image software I've
been able to find requires one to actually look, and perhaps
modify, the bands etc. recognized by the camera. For the
most part, these packages don't export the allele scores to
a text file either - how tedious!
Of course, I think that every autorad should be examined by
the researcher, however, for mapping experiments, I only
care if the genotype has allele A or allele B. Consquently,
I'd like the software to process n number of images, read the
data, then put it all in one file for me.
First, I'd like the software assume an autorad is perfect
(standardized) - see my thoughts on methodology below.
Second, read the tiff file, analyze it, then append output
of genotypic data to a growing data matrix.
The key to automating this process is giving the program near
perfect data ie. no blotches, distinct lanes, clear crisp
bands. I think it could be done easily. See below.
(Anyone done this yet?)
Hand scoring RFLPs:
Create an acetate sheet with grey lines representing lanes.
There should be two lanes more than actually on the gel.
Place the acetate sheet on top of the autorad and color in
the alleles. In the first two extra lanes, artifically color
in the parental genotypes. These will act as approximate
genotype standards for each tier. Then color the bands for each
genotype on the acetate sheet. If there is more than one
allele/probe use two colors.
I place the acetate sheets in a particular order (probe
name), and I write the name in order in a file. I use a
scanner w/ doc feeder to create a separate tiff for every n
acetate sheets (enough for the population size). Every n
scans, the program writes a new tiff with the probe as the
file name. Genotypes (lane id's) will also be read from
a simple text file.
The program then reads and analyses (tiff by tiff) the
colected data. After each analysis, a matrix is assembled
as a simple text file of mapping data.
Anyone have any ideas about working software or peices of
software that would do this? My best lead is Image from
More information about the Bio-soft