The Smith-Waterman (BLITZ) server only does protein-protein searches. If you want to search with a DNA sequence, you must change the "Molecule Type" pull-down menu from 'protein' to 'nucleic'. This will cause your sequence to be translated in all six reading frames and searched against the protein database you have chosen.

It is necessary to input your e-mail address into the "Your email" text-box on the BLITZ server, even if you have chosen 'interactive' in the "Results" box, because the BLITZ server cannot run multiple jobs in parallel. Thus, if you choose 'interactive' and there is a job already running on the server, the result of your job will automatically be returned by e-mail.

You can either type, or cut and paste, your sequence into the large text-entry box. Sequences in GCG, FASTA, EMBL, GenBank, PIR, NBRF or Phylip formats are accepted. If you are using Netscape, you may also upload a file from your computer hard drive, using the "Browse" button.

There are six protein databases that you can choose to search your sequence with.

• swall: non-redundant protein database
• swissprot: SWISS-PROT Protein Database
• swissnew: Updates to SWISS-PROT
• trembl: TREMBL (Translated EMBL)
• tr_new: TREMBLNEW (Translated EMBL Updates)
• swissall: SWISS-PROT + TREMBL + SWISSNEW + TREMBLNEW

There are also several options that determine how easily gaps are created, and how easily they are extended:

• Open Gap Penalty: sets the penalty for opening gaps in the alignments. The default value is 15.
• Extend Gap Penalty: sets the penalty for extending gaps in the alignments. The default value is 1.
• Open QueryGap Penalty: sets the penalty for opening gaps in the query sequence. The default value is 15.
• Extend QueryLen Penalty: sets the penalty for extending gaps in the query sequence. The default is 1.

The "Matrix" pull-down menu lists a number of comparison matrices (BLOSUM, PAM, Gonnet, ID). The default matrix is BLOSUM62. The Gonnet matrix (gb120, gb160, gb200, gb250) series is an updated version of the PAM matrices and are much softer than BLOSUM62 and are thus better suited for finding more divergent matches.

The remaining options deal with the output format, and the scores that are associated with any matches that BLITZ may find.

• "Show Number of Alignments": sets the number of alignments which are returned.
• Show Number of Scores": sets a limit to the number of matches returned.
• "Lowest Z-Score Min List to Report": sets the lowest Z-score to be reported in the result.
• "Show Minimum Quality Score": sets the minimum quality score to be reported in the result.
• "Normalisation Type": sets the method by which Z-scores should be adjusted for length (Default = no normalisation).
  • Log - Uses the log of the length for normalisation.
  • Std - Uses standard normalisation.
  • Stat - Uses Pearson's method of normalisation.
• "Do Averaging of Scores": adjusts quality scores for sequence composition. Default = off.
• Order Output by Overlaps Quality": orders the output by scoring the alignments based on overlap penalty.