If you want to have this newsletter mailed to you or you want to make comments/suggestions about the format/content then send an email to acedb@sanger.ac.uk.
Set your own DNA highlighting colour in FMap, get more information from EMBL dumps and more documentation when you download the acedb server package. There's a new script for setting up acedb databases from scratch, a crop of bug fixes and some the usual nerdy stuff for acedb developers.
The user preferences now include setting your own colour for DNA highlighting in the FMap, this has always by default been a light red/pink, now you can set it to any of the standard acedb colours. Select the "Admin" button on the main window and from there the "Preferences" item.
EBML dumping now includes the following:
The acedb server download package (available from http://www.acedb.org/Software/Downloads/) now includes:
These versions of the files will the correct and up to date versions for the server code downloaded in the package.
(This article is courtesy of Rob Clack rnc@sanger.ac.uk)
(Thanks also to Andrew Menzies who wrote the original version of this script am3@sanger.ac.uk)
Setting up an ace database is not particularly arduous, but there is no reason not to automate it, which is what MakeAceDB does. MakeAceDB is part of AceDB standard release and as such comes as part of the download package available from this site. It is located in the wtools subdirectory, along with all other files referred to in this document.
MakeAceDB is a perl script which creates the necessary directories,
copies the required files into place and initialises the database. If
you specify the path to an ace file, then MakeAceDB will load that into
the new database. The run-time parameters may be given either on the
command line, which does tend to make for rather a long command, or in a
file, in which case you identify the command file with the -f flag.
Examples of how to run MakeAceDB are given in the files MADC an MADF.
There is also an example of a command file, MakeAceDB.com.
You will need to provide the location of a set of .wrm files, but
individual ones can be overridden with local copies if you wish. If
running with command line parameters, just list the .wrm files after the
--tace option. If running with a command file, specify them with
wrm=filename. If your userid doesn't exist in the passwd.wrm file, it
will be added. If you don't specify a passwd.wrm file, a new one will be
created with just your userid in it. You can supply as many .ace files
as you like, and these will be loaded into the new database.
MakeAceDB is written with the intention that you be able to specify different userid and groupid, in which case the userid will be used in the passwd.wrm file, for example.
If you find bugs, please email acedb@sanger.ac.uk.
When the -tsuser command line option was introduced last July no restrictions were enforced on the format of the name used. This led to some errors because users placed characters such as "/" in the name and as this name is used in constructing various file names in acedb this leads to problems because files are created that have the directory separator "/" in their names.
As of this release only alphanumeric characters, i.e. characters from the set [a-z,A-Z,_], are allowed. Hence "my_own_NAME" is valid, but "name with blanks" or "name/with/slashes" are not.
A bug in the GMap cursor code led to the cursor initially not moving when the arrow short cut keys were used. This turned out to be because the code reads a default value for cursor movement from the database and if this was very small, it led to the code calculating that the cursor had not moved. On using the mouse to move the cursor, the cursor movement value got clamped to a minimum value, this now happens when using the arrow short cut keys.
An invalid query sequence before using Dotter caused the dna sequences to be displayed overlapped, fixed now.
Simon gets the prize for finding the oldest bug this month: ArrayInsert() caused a memory overwrite when the array is initially empty.
If you write any bits of code that are common to all acedb main routines then put them in w3/apputils.c.
SMapConvert now uses the new sMapIsReverse() routine to unambiguously find the orientation of single base pair long features. Previously this could not be done as SMap did not carry orientation information separately from coordinate information and since we do not use inter-base coordinates the strand of a single base pair long feature cannot be disambiguated.
As wormbase moves over to using SMap to map more and more of their features there are some changes to the subtags of the Sequence class. For instance, the Allele subtag format has changed from:
Allele ?Allele XREF Sequence UNIQUE Int UNIQUE Int UNIQUE Text
to:
SMap S_Child Allele ?Allele XREF Sequence UNIQUE Int UNIQUE Int #SMap_info
This change of format is likely to happen for other tags as well as we move away from all sequence like features being part of the Sequence class and instead have them smap'd from the Sequence class.
This means that the code must detect this new format but still support the old, the best way to do this is by examining the model for the Sequence class in the current database. The SMapConvert code now examines the model so that it can support the old and new models.
To do this some new calls were put into the bs package (see bs.h and
isOldAlleleModel() function in smapconvert.c for an example)
to allow stepping
through the model without poking into the models BS structs as has been done
on other parts of the code (e.g. treedisp.c or nicedump.c).
We have reached the stage where the strategy is to have constant subparts
(e.g. the SMap tags, the Homol tags) which are then embedded into different
classes, they can all be processed with the same code for virtual sequence building because
the model for the subpart tags remains constant. It makes sense therefore to have
code that can check that these subparts conform to the required format. The
isOldAlleleModel() function was a test to see how this worked, the next
stage is to write a more general function that given a string representing the
expected subtag model will validate it against the model for a given subtag branch
in a particular database.
Rob has been putting a lot of effort into getting the acedb web site more in line with the general Sanger web site in terms of the scripts/Apache directives and general standards used. This is a real improvement because it means we can get more help, more easily from the Sanger web team for web site work.
This has included quite a bit of altering of our basic html files for build results etc. so that they will work with Sanger style Apache like includes etc. He has also managed the move from directly editting the "live" web site to using the internal web development site which can then be copied over to the "live" web site when testing is complete.
He has also put some effort into aquila and finishing off the work started by John Sturdy, we should all be adding new tests to the aquila system.
Sorry no build this month.