Comparative Sequence Analysis
Currently, you have the following options:
To initiate a SMART analysis, first you have to cut and paste your query protein sequence into the text area (do not paste in header information).
Please note that the minimum sequence length supported is 50 amino acids, and DNA sequences are not supported.
You may restrict the search to intracellular or extracellular domains.
In addition, the user may view the result of a WU-BLAST search of the domain sequence database with the query sequence (here, just click on the corresponding checkbox).
Including the PFAM analysis, the SMART program initiates a search of the Pfam domain database (using the HMMer program hmmfs, and thresholds determined by Pfam contributors).
Including the additional HMM searches, SMART initiates a search of the SMART domain database using either of the hmmfs (for alignments across at least part of the HMM) or hmmls (for alignments that span the entire HMM) HMMer programs and a user-supplied bits score threshold between 5 and 55 bits (in steps of 5)
Hit the "SMART" button to submit the request.
SMART searches your query sequence for the occurrence of characteristic domains. Upon success, these domains are represented graphically as "bubbles" on a line.
If SMART predicts no domain, you may want to restart your search with a less restrictive threshold. Using the selector, the threshold can be lowered by 1, 2 or 3%.
Predicted signal peptides, transmembrane segments, coiled coil regions and segments of low compositional complexity are also represented in an extra line of output.
What is shown is similar to the example given below:
Note that many things in this output are "clickable": annotation, alignment, consensus terms and further analysis information can be found by clicking on the bubbles, or else the lines themselves. See "Domain Prediction Output" for further information concerning all such possibilities.
Perform a SMART analysis on a query sequence indicated by its sequence identifier or its accession number. Only those proteins containing domains represented in the SMART collection are able to be retrieved.
Sequences can also be specified using their sequence identifier (ID) or their sequence accession code (ACC).
These codes/identifiers may be drawn from all popular databases (including SwissProt, EMBL, TrEMBL, PIR, and PDB). Note that GenBank identifers [gi numbers] are not (yet) supported.
Multiple sequences that have similar accession codes (e.g. UROK_HUMAN, UROK_CHICK etc.) can be searched using a truncated version of the code (i.e. UROK). Only those proteins containing domains represented in the SMART collection are able to be retrieved.
If you initiate SMART using the ID or accession number options, the output shown depends on the number of hits matching your query.
If the complete ID or accession number is specified, the output is identical to that described above.
If an incomplete ID or accession number is supplied then SMART will first present you with a list of all sequences that matched your query (see the Figure below).
Up to 25 sequences from this list may be selected to restart SMART, which will then display all predicted domains as 'bubble pictures'. Note, however, that selecting multiple sequences from this list doesn't allow output of additional predictions such as transmembrane segments or coiled coil regions, and nor does it allow WU-BLAST2 searches.
View the annotation pages of domains included in the SMART set
The SMART Homepage contains a list of those domains that are able to be predicted currently.
Clicking on the name of any one of these leads to that domain's "Annotation Page" containing information relating to sequence, structure and function.
