by Roderic Guigó 5/6/97

In this practical we will run a sequence against PFAM, a database by Ewan Birney, of protein domains. In order to characterize the function of an unknow protein, searches against databases of motifs corresponding to protein families may be more sensitive than simple sequence searches. PFAM is a database of protein domains represented as Hidden Markov Models.

WWW DB Tools

We will use:

• SRS (at EBI or EMBL) to extract a query sequence.
• PFAM ( at the Sanger Center

Step 1. Extract an hypothetical protein from the database

• Load SRS (at EBI or EMBL) and Start the session.
• Select the Swiss-Prot database and continue.
• Typehypothetical in the Description box, and similar in Commentthen Do Query.
• Clik on Y4MN_RHISN entry (at the bottom of the first list of 30 hits).
• clik on save
• Sequence format is fasta and view FastSeqs. Then click on save.
• Netscape (text) save on e.g. hypothetical.fa

Step 2. Compare the sequence of the hypothetical protein with Pfam domain database

• Load PFAM Server
• With Cut and Paste the sequence of the hypothetical protein
• Now Search Pfam W (it may take a while)
• Analyize the results. Try java viewer.
• Click on transketolase domain.
• get alignment (of transketolases).