In article <316 at uncmed.med.unc.edu>, danielg at uncmed.med.unc.edu writes:
> Am posting for a friend. He needs a program that he can run on his pc to
> estimate band sizes on Southern blots. His description was sketchy, bear
> with me.
>> Currently, he uses a prg written in quick basic, in conjunction with a
> digitizing pad. His program works like this. Put the southern blot on
> the pad (visualization done by biotin labelling), and then he touches the
> pointer to the origin in lane #1, then on each band in lane #1, which is
> MW markers. He enters the band sizes, and the prg figures a std curve,
> basically a plot of fragment size vs. migration distance. From this, he
> can estimate band sizes in the following lanes.
>> His problem is that he loses accuracy with fragments larger than 7 kb.
The problem with loss of accuracy at > 7kb can't be solved by any program -
it is inherent in the nature of gel electrophoresis. A solution is to run
gels with lower agarose concentration for cases where large fragments need
to be handled. Then it depends on how large large is. Pulse field gel
electrophoresis is needed for really large pieces up to whole chromosomes.
As you know computers can't be more accurate than the input data - and that
is the problem here.