Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!psuvax1!psuvm!MITWIBR!GILBERT From: GILBERT@MITWIBR.BITNET Newsgroups: bit.listserv.info-gcg Subject: (no subject given) Message-ID: Date: 7 Feb 90 15:41:00 GMT Sender: "INFO-GCG: Univ. of Wisconsin Genetics Analysis Software Discussion" Reply-To: GILBERT@MITWIBR Lines: 80 Approved: NETNEWS@PSUVM Gateway Original_To: JNET%"info-GCG@utoronto" I've been following the GCG interface discussion with great interest and would like to describe for all of you what we have done/are doing here at the Whitehead Institute regarding the GCG interface. Mike Cherry's note actually tipped my hand a bit when he referenced the development effort using Apple's MacWorkStation with GCG on the VAX. First of all, the the front end work we have done works only with the GCG programs which first ask for a sequence(s) and then ask for a set of parameters which then run to produce a file or plot as output. The front ends do not work with programs such as REPEAT, SEQED or LINEUP. For the command line types the GCG package is available in it's "native" form. I think in the last 3 years we've had one person who wanted to use it this way. No flames please, I'm just reporting the facts. The major mechanism which we use is very simple. Let me describe.. For each GCG program we have written a short (2 page) C program which asks the users for all the input parameters for the target program. We are able to control the wording of each prompt and that is a major help. We also put alot of thought into the defaults which will be used. Even though our "version" of MAP may ask over a dozen questions the users learn quickly to race through the questions. Further, I also have never had a user have to ask how to do a map with silent mutations, or how to restict the range of the number of cuts. For our site, asking all of the questions works. We also combine several GCG programs into one front-end, for example COMPARE & DOTPLOT, WORDSEARCH & SEGMENTS to name a few obvious combinations. As an aside our front-end programs do their own sequence retreival so we are able to use the PIR formated databases directly off tape. Our users can also move sequences from their personal computers to the VAX and use a qualifier on the filename to have the program read the format ie. MYSEQ.SEQ/IBI will read the file MYSEQ.SEQ and know that it is in the IBI (GenBank) format. Other supported formats include: /RAW for sequence only, /STRIDER for DNA Strider, /IG for you know who, /NBRF and /GCG. Globaly, we can set a default for what the format should be if no qualifer is given, here we use NBRF other place where I put this software up use GCG as the default. The front-end programs also keep track of name of sequence currently being analyzed and they pass that name from program to program, so one can do a MAP,MAPSORT,MAPPLOT and FASTA by typing the sequence name once. It then becomes a default for subsequent programs. Finally, as Mike Cherry mentioned we are developing a MacWorkStation application to serve as an interface to the VAX and the GCG package. For those unfamiliar with MWS here's a brief description. You have an MWS document on your MAC which contains a script for navigating the connection from your MAC to your VAX. The simplest case is where you are directly connected via an RS-232 wire, in a more complex case the script would have to look for and dial a modem. Either way the end-users doesn't see any of this. Once connected to the VAX, a VAX program is run which conmmunicates with the MWS program on the MAC. All of the programming is done on the VAX so you don't have to learn to program on the MAC. The VAX language may be anything you like, here we use "C", but I digres too much on implementation details. The VAX program then builds menus, and waits for the MAC to respond with menu choices made by the user. If the choice "Restriction Map" is selected from the menu the VAX puts up a dialog box which contains all of the MAP/MAPSORT/MAPPLOT options, all of which are already filled in, so most of the time the user presses the OK button or a carriage return. The VAX program then reads the contents of the dialog box, determines if a MAP, MAPSORT or MAPPLOT has been requested. Run the GCG program for the user and move the outfile into a MAC text window. At this point the user may print the window, review and edit the window, cut/paste to another application, save the window to either a MAC or VAX file or simply delete the window. When the MWS application is Quit, the VAX program shuts down and logs off. During all of this the user has only a faint idea that they are even attached to a Mainframe, the wait times are longer for program execution. We allow the users access to their mainframe files via a "Host Files" menu which displays their files in a window using "View by name" style used by Finder. Even here we show files size in "K" instead of blocks and use a MAC style date. Renaming, copying, typing, and deleting are all done from the "Host files" menu, as are any "SET DEFAULT"'s which are implemented as "Host folders". All in all, it is quite robust and hides the mainframe very effectively. The work describe above is avaliable for export but I'ld prefer not too at this time. I am much too busy, as I hope you can tell, to support lots of other sites. I'ld be gald to shovel over what I have but make no promises about it's applicabilty to a target site. The preceding discussion was intended as a "Where we are and what we done" talk. Let here from some other folks out there. Will Gilbert, Ph.D. Whitehead Institute Gilbert@MITWIBR.Bitnet