Gap stores the data for each sequencing project (e.g. the data for a single cosmid or BAC) in a gap assembly database, so at the start of a sequencing project the user should employ gap to create the database for the project (see section Opening a New Database)
Gel reading data in experiment file format (see section Experiment File Format) is entered into the database using the methods available from the assembly menu (see section Entering Readings into the Database (Assembly))
To assemble more data for the project or to edit or analyse readings already entered the user should open the same project database (see section Opening an Existing Database)
Although the database files are designed to be free of corruption it is advisable to make regular backups (see section Making Backups of Databases)
Database names can have from one to 240 letters and must not include a full stop or spaces. The database itself consists of two files; a file of records and an index file. If the database is called `FRED' then version 0 of the database comprises the pair of files named `FRED.0' and `FRED.0.aux', the latter of these being the index file. The version is the character after the full stop in these filenames. Versions are not limited to numbers alone, but can only be one character.
When a database is in use a `BUSY' file is created. For the
`FRED' database this will be named `FRED.0.BUSY'. The file is
used as a safeguard against two people editing the same database
simultaneously. In the event of a program or system crash this busy file
will be left on the disk, even though the database is not being used. In
this case remove the busy file (after checking that it really isn't in
use) using the unix
rm command before opening the database. Eg
The earlier assembly program,
Bap, used a different database format
consisting of five files.
Gap cannot access these old databases
directly. We provide a convert program (named appropriately enough
convert) to create gap format databases from either bap or dap
databases. See section Converting Old Databases.
The gap database is robustly designed. Killing the program whilst updating the database should never yield an inconsistent state. A "roll-back" mechanism is utilised to undo any partially written updates and revert to the last consistent database. Hence quitting abnormally may result in the loss of some data. Always quit using the Exit command within the File menu.
However it is advised that copies of the database still be made regularly to safeguard against any software bugs or system disk corruptions (hopefully both will be unlikely).