Version-1997.0 Release Notes

Some of the changes in this release will not be obvious to the user. Many changes were required to improve the Gap4 scripting language. We consider this important as it puts us in a better position for future development and allows other people to easily extend gap4 themselves. The first draft of the documentation on this scripting system is now available and can be dound in the scripting_manual/scripting_toc.html file.

The most noticable changes are within gap4, however two new programs have also been added. Vector_clip is is a replacement for vepe, which is still distributed at present. Trace_diff is a tool for comparing pairs of trace files to produce a "difference" trace file and to automatically find and mark mutations.

Gap4 has a much improved Template Display that can now display more than one contig. This provides a much easier way of investigating the relationships between contigs.

Gap4 editing has had several improvements too: it redisplays faster; superedit now provides a larger diversity of editing abilities; a vertical scrollbar, trace differencing, and showing disagreements by colour aid using the editor for mutation studies; and the disassemble readings and break contig functions now have interfaces available directly from within the editor.

Trev has had a few updates to allow repositioning of the vector clips. We've also fixed a couple of bugs to do with saving the accuracy values after editing.

Pregap is now more robust and has an extra question to control whether it should derive reading names from ABI samplenames (held within the files themselves) or the ABI filenames.

Dynamic linking is now used for many programs. This should be transparent to the user, but should use fewer system resources on crowded machines and allows for greater flexibility. The programs using tcl and tk have been updated to use tcl version 7.6 and tk version 4.2 to allow for on-the-fly loading of dynamic libraries, which translate to "packages" to extend gap4, sip4 or trev. See the scripting documentation for details.

Finally, trace file compression is better supported. Gap4 and trev can compress and uncompress trace files (whatever format) on the fly, thus allowing for significant reductions in disk space. The "gzip_trace" and "bzip_trace" scripts can be used to compress your SCF files, saving around 70% of your disk space, which is equivalent to at least a 10 to 1 compression of ABI files.

A more verbose list of changes follows. Note though that only the more important bug fixes are listed, and we consider those marked with asterisks to be most important.

New Programs

• Vector_clip program to replace vepe. Also used by pregap.

• Trace_diff program to compare a two traces to find mutations.

Gap4

• The Template Display now supports displaying multiple contigs including displaying their relative positions and overlapping data. It also allows for their relative positions to be changed.

• Major revamp of the gap4 scripting language including the first draft of the scripting documentation.

• Shuffle pads now removes places where there are columns of 100% pad.

• The "align" button in the join editor now adjusts the relative offsets between contigs when the alignment requires it. It used to insert unnecessary pads instead.

• Improved superedit to allow user control over which individual editing functions are activated. Includes insertion to reads, deletion to reads, insert of anything to consensus, deletion of dash in consensus, deletion of anything in consensus, replacing of bases in consensus, shifting readings, transposing any two characters (usually only allows pads), and allowing uppercase for changes/insertions. Each of these options can be independently controlled.

• Maximum database filename length is now 256. NB: this is not the reading name length.

• Vertical scrollbar in contig editor. Used when the height is greater than a specified threshold.

• Use of tcl7.6/tk4.2 (instead of 7.4/4.0). We utilise the tcl dynamic loading strategy to allow arbitrary user defined extensions to be added directly onto the main menus.

• Improved handling of read only databases, including greying out of menu items and a read-only editor mode.

• Bug fix: Suggest Primers when only using one contig now works.

• Bug fix: When assembling really long (4K) readings we now check that adding pads doesn't overflow the maximum allowed length.

• Directed assembly has improved memory usage checking. Better checking of the tolerance used during assembly.

• Better low level error checking and reporting.

• Sped up consensus calculation (by upto 50%).

• Disassembly and break contig from within the contig editor.

• Bug fix: Tag positioning during assembly was sometimes slightly out, due to padding sequences.

• Bug fix: Shuffle pads could occasionally completely corrupt the database.

• Automatic 'maxdb' parameter adjustment when opening large databases. However it still needs to be hand set for large assembly runs.

• Bug fix: Fixed rare crash in trace display.

• Bug fix: Searching by tag type and annotation (in the editor) now searches in order of the left end of the visible region. Previously we searched on the left end, which caused some tags to be missed when they overlapped between used and cutoff data.

• Bug fix: Doctor database was reporting the 'referenced by' field of annotations as one reading too high.

• Melting temperature is now listed in the suggest probe output.

• Bug fix: Doctor database's shift readings function works once again.

• Added a "read pairs" only display mode of the template display to only show templates containing both forward and reverse readings.

• Added a Print List command in the list menu. Printing is also possible from the main output window.

• The editor "set output list" setting no longer clears the list. This simplifies producing lists from several contigs.

• The join editor can now be brought up from the read pair plot.

• New control-T key binding in the editor to display traces.

• New "delete from left" command in the editor.

• Bug fix: Removed memory corruption in the primer selection code (both in the editor and automatic modes).

• Bug fix: Removed memory corruption in check assembly.

• Bug fix: Removed bad memory free in find oligos.

• Extract Readings can now ouput directed assembly data. Coupled with Directed Assembly this provides an efficient and quick way of splitting and merging databases.

• Bug fix: Directed Assembly can now correctly handle readings containing consensus tags.

• Improved the list mechanism for auto-generated lists. A list named "[contig_identifier]" will expand to all the readings within a contig. A list named "{list_name}" will expand to all the readings within all the contigs in a list (eg "{contigs}"). Additionally "allreadings" and "allcontigs" lists are calculated on the fly.

• Double left clicking on a match in the 2D plot now performs the "obvious" operation (which is the one marked with an asterisk in the popup menu).

• Check assembly no longer brings up new editors when one exists, but moves its displayed position instead.

• Contig Editor speed and X network traffic usage has been improved, specifically when editing single sequences or displaying very colour rich screens.

• Trace differencing added to the contig editor. We now have the ability to compare any two traces to display differences, or to compare any trace with a 'standard' comparator.

• Bug fix: solved editor crashes where the quit button was pressed either when busy (eg aligning) or just starting up.

• Editor reading name selection now updates the global 'readings' list.

• Bug fix: contig or reading lists containing duplicate entries no longer cause problems.

• Removed unnecessary template display and contig selector redisplays whilst double stranding and suggest primer routines are running.

• Bug fix: it's no longer possible to edit 'deleted' annotations in the editor.

• The quality plot can now be display independent of the template display.

• The contig editor can now highlight differences by colour in addition to the previous 'by dots' method.

• Bug fix to list names listed in the 'list' name browser.

• Renamed Save As as Copy database. Garbage collection is now enabled by default.

• The maximum length of trace filename is now 256 (was 32).

• Updated show relationships to take a list or a file of contig names.

• The "Information" option from popup menus (contig selector, template display, etc) now also brings up a dialogue box.

• Can now bring up the editor directly from the stop codons, restriction enzymes, and quality plots.

Trev

• Major update the the tcl code to use the new menu system and official tk filebrowser.

• Saving SCF files now correctly updates the confidence values.

• Added a font menu.

• The AV line is now output to experiment files when editing in trev.

• Trev now supports setting the left and right vector cutoffs.

Pregap

• Program failures (eg eba) can no longer cause the "out of sync" problems where names cycled around.

• Added checks for incorrect (based on number of words) experiment file line type values during the augment stage.

• Fixed possible echo '\c' problem on Solaris 2.x systems.

• Now uses the vector_clip program instead of vepe.

• New question to handle cases where ABI files have incorrect or missing sample names.

Sip4

• Bug fix: Allow larger number of hits in the library searching code.

Xgap

• Bug fix: Removed the disk space deallocation system (added during last release) as it proved to be only compatible with gap4. Consequences were completely corrupted databases (rare).

• Bug fix: The contig editor now correctly flushes changes when saving. Previously consensus tags edits weren't saved until the next command saved.

• Added missing files from previous release. Specifically these are BAP.MRG (used by gap/xgap) and some bap source files needed for compilation of gap/xgap.

Misc

• Changed fonts for some dialogues. This solves problems for some Solaris systems.

• ALF reading code now supports traces with bases at negative positions. We simply reset them to position 0.

• The clip program now sets QL as well as QR. It also supports minimum and maximum allowed values.

• Copy_db now generates consistent contig order information when merging multiple databases.

• Added an error_bell command to the tcl/tk programs to turn off ringing the bell when outputting errors. Useful within scripts.

• Improved compressed trace file support for trev, gap4, eba, makeSCF, etc.

• Added a WT (wild type trace) experiment file line type.

For bug fixes to this release please see the news.230497a.html and news.070597a.html news pages.