Can you help?

Michael Myers myersm at rockvax.rockefeller.edu
Thu Jul 27 11:52:04 EST 1995


In article <3v386h$cnk at dartvax.dartmouth.edu>, Bob.Gross at dartmouth.edu
(Robert H. Gross) wrote:

>Damian -
>
>There is a program called the Gene Construction Kit which seems like it
>will meet your needs. A demo is availabe by FTP from most Info-Mac
>sites and from Don Gilbert's site at Indiana. You can also contact the
>company (Textco) directly at textco at valley.net.
>

I apologize that the following is so long. I got a bit carried away!
Continue only if you're interested in my personal opinions regarding GCK,
MacPlasmap, and GCG.

I have used Gene Construction Kit a fair amount. First, you should know
it's fairly expensive. I believe nearly $1000, make sure you ask about an
academic discount. Secondly, while some of its features are quite nice, it
has some annoying qualities. For example

1) the postscript output of circular maps is not that great. The arcs with
arrow heads come out misshapen. The output from MacPlasmap is much better
in my opinion. I feel this *really* reflects poorly for a program of its
price and purpose: high quality graphic output of plasmid maps. I can
produce better looking maps by doing it manually in Canvas. Also,
automatic placement of restriction sites and ORF labels on circular maps
needs to be greatly improved. For example, if you've given a certain
region of the construct a name such as lacZ or bla, GCK cannot put a text
label for that name next to the proper arc in the circular map. Again,
MacPlasmap does a better job of this. It puts it there, and if it's
crowded with other text labels, you just grab it move it! Sure you can add
text in GCK, but you have to go the illustration window, and the label is
not specifically linked to its corresponding region of the construct.

2) you have to pick precisely which restriction enzymes you want
represented on your map (ie. you cannot do as in GCG: show all enzymes for
6-base and higher recognition sites, with a maximum cut number of N).
Neither does it show enzymes excluded or enzymes that don't cut. In short,
GCK does not fulfill both chores of restriction mapping and plasmid
drawing. To get better restriction mapping, you have to spend a lot more
money and buy Textco's program (soon to be released) called DNA Inspector
(which includes a myriad of DNA/Protein analyses).

3) you have great control over formatting of a nucleotide/protein
sequence. Any segment can be selected and its formatting (font, size,
style, color, group size) adjusted. Overall this works quite nicely.
Except: the amino acid translation for a corresponding DNA sequence can
only be shown in three letter format, not single letter. This also means
that you must use a monospaced font (monaco or courier), otherwise the
three letters of each amino acid are not kerned properly (get extra space
between some letters). Also, there's no way to tell the software to splice
ORFs togther when doing the protein translation, which screws things up if
you want to have introns present (the amino acid numbering is off). For a
sequence we are about to publish, I ended up using GCG's "publish" program
to make a text file with the DNA and protein sequence, then imported it
into Canvas for additional formatting.

4) GCK uses "lists" to store nucleotide sequences such as the restriction
enzymes and their sites, or primers that you routinely use. Thus for the
latter, you can pick your primer sequences from a list you created, and
they will be displayed in the nucleotide sequence. One missing feature:
the ability to import lists! I have about 40-50 primer sequences that I
would have to type in by hand if I want to build a GCK list. It would be
really simple to add this feature to the program. Primers of your own
making are the most effective landmarks in nucleotide sequence, even
better than restriction sites. (Don't forget that you can add primer
sequences and their names to the enzyme.dat file that GCG uses to map a
DNA sequence, which means their names show up in the output, just like a
restriction site.) 

5) GCK's method of setting page size (margins) is really kooky. Still
haven't quite figured it out. Trial and error is the only way. 

6) A feature that I like a lot: you can have GCK show the nucleotide
number of the cut site in parentheses after the cut site label, eg.   
NcoI (2345).  This is really convenient when you're subcloning and you
want to quickly determine fragment sizes for analyzing gels. With the
output from GCG's "map" program, you constantly have to read off the "tick
ruler". Of course, when you cut and paste DNA segments together in GCK,
the numbering gets updated. 

Here at the Rockefeller, we have a licensed version of MacPlasmap running
on the network with KeyServer. This permits a certain maximum number of
users at any time to run the program from a server. Works pretty well and
it produces very nice laserprinter output. I'm not sure how much this type
of licensing arrangement costs.

To summarize: I like a lot of GCK's features (to be fair, there are many
good features I didn't describe here) , but for the price that Textco
charges, there are some glaring holes in their program. I suppose the most
telling assessment is that I still use the mapping programs in GCG! (By
the way, I highly recommend Canvas for other aspects of making figures.)

OK, now flame me!

-- 
Michael Myers
The Laboratory of Genetics
The Rockefeller University
myersm at rockvax.rockefeller.edu




More information about the Bio-soft mailing list