Database Operation

Spotfinder can upload the results of image analysis directly to the TIGR compliant microarray database. However, user must submit experiment info using MADAM all the way through Scan Page before to start any data submission from Spotfinder directly or from MEV files. This is necessary because Spotfinder uses the loaded tiff image file names for database submission. The loaded files have to have the same names as files submitted on Scan Page of MADAM.

Actual data upload from Spotfinder to database is performed with the help of a program called a SpotfinderDBclient. SpotfinderDBclient is written in java and can be launched locally or from any location on network, where it is accessible from local machine. By default SpotfinderDBclient is installed in the same directry as Spotfinder itself. SpotfinderDBclient uses jdbc to connect to database and requires java runtime environment (JRE) to be istalled. Distribution comes with two jdbc driver files: one for SYBASE, and another one for MYSQL database. Before to start data submission to microarray database user should launch Database Settings Dialog from Settings menu to make appropriate settings. User would need to select database type (SYBASE or MYSQL), server name, database name, user name, and password. Go to Database Settings Dialog for details.

To start database submission user should go to Data on menu bar and select command Upload results to DB. This function creates temporary mev file in the same folder where SpotfinderDBclient is located. The name of temporary mev file is based on the tiff file name currently loaded in channel A. Besides this the function creates DBlog.txt file, if it does not exist yet, or update it, if it is located in the same folder as SpotfinderDBclient. DBlog file is tab delimited text file and is used by the SpotfinderDbClient. The DBlog file keeps records of all submitted jobs: active, pending, and completed. User may rename or move DBlog file to another location when it becomes to big in size. DBlog file has five columns: analysis ID, temp mev file name, tiff filename in channel A, tiff filename in channel B, and comments. When user selects Upload results to DB one line is added to DBlog file that sets analysis ID to zero and stores all needed file names. The comment line default string is set to "not processed". After that Spotfinder launches SpotfinderDBclient, which in its turn reads DBlog file and processes data line by line. If during the data submission to database user send another data set to DB from Spotfinder, new line of records is added to DBlog file. The new record will be processed after the previous one is completed.

When submission is completed the active record is changed. Actual analysis ID is set in the first column and the information about the number of rows submitted to the expression table on selected database is stored in comments section. If for some reason the data submission to database failed the analysis ID remains set to zero. That means that next time when user starts DB submission SpotfinderDBclient will read unprocessed line from DBlog file and try to submit it to database. In this way DBlog file can work as a dispatcher for batch database submission.