[Molecular-evolution] Re: GCG non-support

Peter Rice via mol-evol%40net.bio.net (by pmr from ebi.ac.uk)
Thu Aug 2 05:15:04 EST 2007


I just saw a messsage in this thread in the bionet.software and
bionet.software.gcg newsgroups (they will match the bio-soft and info-gcg
mailing lists I noted in the header).

I fear I may be missing other messages.

Steve Thompson wrote:
> I too am dismayed at Accelrys' terrible plan to 'retire' GCG.  Yes, they 
> argue all of the individual components are available in the public domain, 
> and sure EMBOSS even puts most of them under one umbrella. However, and I've
>  been looking for years, I completely agree with your letter (included below
>  in its entirety) there is no GUI/sequence editor out there that comes 
> anywhere near approaching SeqLab's functionality.

When Alan Bleasby and I started EMBOSS some 11 years ago, we deliberately
avoided writing a GUI in the hope that someone, somewhere would come up with a
killer GUI and we could just fit under it. Although there have now been over 100
interfaces and packages developed out there that include EMBOSS, none seems to
have taken over.

One of the interfaces was indeed SeqLab which one of our commercial users
reported customising to include EMBOSS some years ago. As far as I am aware it
was only ever used internally.

> Regardless, I have personally urged my rep' there to encourage the 
> 'powers-to-be' of allowing the package to move on elsewhere, even if it's 
> just SeqLab's source, and preferably to the public-domain.  If Accelrys has 
> decided they are not making enough money on it, then let it continue 
> elsewhere, so that its 25 year legacy of making sequence analysis 
> approachable to scientists worldwide can continue! Furthermore, I am willing
>  to organize, and wish to begin, some type of a petition process that will 
> allow scientists over the world to add their voice to this opinion.  Since 
> this is such recent news I have not yet began this process, but am anxious to
>  get started, and happy to entertain others' thoughts on the matter.  Let's 
> not let it silently fade away.

SeqLab was originally GDE, written by Steve Smith at Harvard and much extended
when he was at GCG Inc.

If SeqLab does appear as public domain or open source, then I may be able to
help put EMBOSS under it. Many years ago (before EMBOSS) I modified SeqLab as an
interface to its predecessor the EGCG package. The code was pretty clean and it
was not hard to do.

Making SeqLab's source public domain (or open source) is a little tricky as it
depends on things like how GDE was licensed to GCG Inc and who contributed in
the past.

The licence in the 2.2 distribution (last release before GCG adopted it) is a
little unusual by today's standards. The GDE.readme file says:

   The programs are not in the public domain, but are and will continue to be
   available for free.  Distribution includes full source code, and binaries
   along with a users manual.

That was from 1992. Is it perhaps an early open source licence? I read "not in
the public domain" to say "author retains copyright" but I see no restrictions on
modification or reuse. There is also no obligation for later modifications to be
under the same licence.

Release 2.2 is still available from
http://iubio.bio.indiana.edu/soft/molbio/unix/GDE/

I have seen a revival of GDE, but presumably not with the full functionality
that was added to SeqLab. See http://www.msu.edu/~lintone/macgde/

The MacGDE page says they believe GDE cannot be distributed by commercial
enterprises, but I don't see any prohibition. The licence is the original GDE
one. I suspect BioLateral were just being cautious.

> On Wed, 1 Aug 2007, Staffa, Nick (NIH/NIEHS) wrote:
> 
>> I am the employee of a sub-contractor serving in IT support at the National
>>  Institute of Environmental Health Sciences in Research Triangle Park, NC. 
>> I have supported GCG and sequence analysis at this Institute since before 
>> GCG even had fragment assembly.

Ah, the old crowd are still around. I first encountered GCG in 1987 at EMBL
Heidelberg when they were still "UWGCG" and was one of the first users of their
fragment assembly. Those were the days! (yikes! ... they were 20 years ago :-)

>> Accelrys's plan to "retire GCG products" comes as rather a shock. They were
>>  right in the middle of bringing it into the genomic age. It is especially
>>  alarming since many IT types seem to want to get rid of un-supported 
>> software. NIH in Bethesda MD has removed GCG and urges people to use 
>> EMBOSS.

I have not seen any announcement (though I was aware of the rumour).

Two questions:

Does it say what will happen to the GCG products after retirement, or is that
still to be decided?

Does the GCG licence still allow you to run GCG after it expires? (Or, to be
more specific, running it on your own machines, for external access through
SeqWeb, for logins from outside)

>> I don't think EMBOSS - an eclectic collection of miscellaneous programs - 
>> approaches the usefulness of the SeqLab environment. One very useful and 
>> important feature of SeqLab is its ability to easily make lists and ad-hoc
>>  databases (with LookUp and other programs) and to do fasta or other 
>> searches in them.

Three separate issues here:

EMBOSS functionality can be compared to the GCG applications,
and we do hear from our users that most things in GCG are covered
(peptidestructure/plotstructure is an outstanding request, RNA folding is
covered by the ViennaRNA package as an "EMBASSY" addon to EMBOSS, paup is
covered similarly by PHYLIP.

LookUp I never liked ... it is SRS version 4, only partly integrated (database
links and other features missing). SRS version 5 was still public but they never
made the (huge) effort to integrate it. I always had SRS available locally
(it came from EMBL Heidelberg) so I always turned GCG's LookUp off. As I have
worked with SRS rather a lot, including a couple of years at LION Bioscience
writing parts of it, I'll shut up about it :-)

SeqLab is a big issue - we do not know of an EMBOSS interface that quite
compares (and I rather miss GDE)

>> GCG started out at the University of Wisconsin as a collection of existing
>>  programs that were given a unified (for want of a better word) user 
>> interface by the then UWGCG.  There was a hint that the founders were 
>> thinking of a consortium of some sort - a cooperation among users, 
>> scientists and developers. The sources were available up through version 8.

Indeed, I was very much part of that community :-)

When EMBOSS started I had to kill EGCG. GCG claimed rights to the code and
refused to allow its reuse, so I claimed rights to the code and refused to allow
them to continue distribution. All the useful applications had to be rewritten
from scratch for EMBOSS. Only prettybox survived - Rick Westerman contributed it
to GCG.

There was a release 9.0 of EGCG (despite the GCG source code licence problems)
but it only ever went for beta test at one institute, and that institute is now
closed down.

I wonder what the future holds ...

Peter Rice
European Bioinformatics Institute








More information about the Mol-evol mailing list