firestar[1] makes predictions of functionally important residues using the large inventory of functionally important residues in the FireDB database [2]. The reliability of the transfer of functional information between the functionally important residues in FireDB and the query sequence is evaluated via the local residue conservation between the two sequences.

User Input

• PDB code: each entry in the PDB is identified by a four letter code (e.g. "1tco"). PDB files usually contain more than one chain, so the user is prompted to give the chain identifier.
• Uploading a PDB formatted coordinates file: Here again the user is prompted to choose the chain if more than one appears in the file.
• Sequence input: At present we only accept FASTA or raw sequence formats. In this case some options will be disabled because no structure can be used for structural alignments.

Output

Target sequences are subjected to standard PSI-BLAST searches. Profiles are generated with an nrdb90 database from the EBI and the final search is made against the FireDB consensus sequence database derived from PDB sequences.

E-values can be modified by the user. The default E-value cutoff is set to 10. We have found [3] that in some cases local similarities are reliable enough to transfer binding site residues between even distant homologues.

Only those PSI-BLAST hits with any functional information annotated in FireDB are displayed. Each hit is represented in its own box. All boxes are displayed on the same page.

Each alignment is evaluated using a specially adapted version of SQUARE [4]. This method generates scores for each position in the alignment based on local conservation scores.

SQUARE produces reliability scores for all aligned residues, but it is important to note that these reliability scores refer solely to the viability of the transference of functionally important residues from the consensus sequence. The score is meaningless for those residues in the FireDB consensus sequence that do not have associated functionally important residues.

The functionally important residues from the FireDB consensus sequences are supported by evidence. In the case of catalytic residues the evidence comes from the Catalytic Site Atlas [5] and are annotated as "literature" or "inferred with PSI-BLAST". In the case of binding residues a percentage of occurrence is generated. The following table explains evidence and reliability scores in more detail:

Structural alignment output

Structural alignments can be generated by clicking on "run LGA" in the yellow box next to each PSI-BLAST output.

Structural superpositions between the query and the representative of the selected cluster are generated with LGA [6] and open in a separate page that has a similar look to the pairwise PSI-BLAST results.

The superpositions can be visualised within the page with a Jmol applet. Note that this means that Java environment must be installed for the navigator. If the java environment is not installed for the browser the applet may cause the navigator to crash on some platforms. Unfortunately this technical issue is beyond our control.

Multiple sequence alignment output

Multiple alignments between the consensus sequences and the query sequence can be generated by clicking on the "Display Multiple Sequence" button in the pink box in the PSI-BLAST output page. MUSCLE [7] will align the query sequence to all the consensus sequences found by PSI-BLAST.
The MSA output page allows selection of a subset of sequences to be realigned by checking in the checkboxes.

Firecat help page

Alignment reliability is most sensitive to alignment quality. This tool allows user to insert their own pairwise query-template alignments with the requirement that template must be a PDB sequence.

It requires the second sequence in the pairwise alignment to be a PDB sequence. In this way the server can identify the FireDB cluster to which the second sequence belongs and associate the functional information stored in FireDB.

This tool can be used when alignments can be improved, but also when firestar PSI-BLAST is unable to find homologues. In this case query-template alignments from fold recognition servers can be used.

example

The following alignment has been generated by the FFAS03 [8] threading server and was the best scoring alignment for target T0369 of the CASP7 experiment in the 3d-jury [9] metaserver. The structural template identified by FFAS was 1rxq_A and we can assess whether the functional residues transfer between the target and the template.

>T0369
MTDWQQALDRHVGVGVRTTRDLIRLIQPEDWDKRPISGKRSVYEVAVHLAVLLEADLRIATG-------- 
-----------ATADEMAQFYAVPVLPEQLVDRLDQSWQYYQDRLMADFSTETTYWGVTDSTTGWLLEAA 
VHLYHHRSQLLDYLNLLGYDIKLDLFE
>1rxq_A
SKEQKDKWIQVLEEVPAKLKQAVEVMTDSQLDTPYRDGGWTVRQVVHHLADSHMNSYIRFKLSLTEETPA
IRPYDEKAWSELKDSKTADPSGSLALLQELHGRWTALLRTLTDQQFKRGFYHPD-TKEIITLENALGLYV 
WHSHHHIAHITELSRRMGWS-------

We paste the alignment in the text box, with the template PDB code: "1rxq" and the chain identifier: "A". And the firecat output is:

Based on this output, the nickel binding site can be reliably transferred from 1rxq_A to the target T0369. Note also that 1rxqA could not be obtained by standard PSI-Blast searches.

Evolutionary relationships between sites

Evolutionary relationships guide scheme