SEDB has two main search criteria, the first search is based on structural similarity and the second search is based on sequence similarity. When performing a structural search, SEDB accepts a PDB code, chain identifier, RMSD and Z-score cutoffs to create a table to display structurally similar proteins (A1). For each structural neighbor in this table the related sequences with exon data are collected from local EID and presented in the table. To speed up the data retrieval, the alignments between PDB structures and EID (GenBank) sequences with annotated gene structure have been pre-calculated using BLAST. The following thresholds have been applied: sequence similarity at least 90% and alignment length at least 50% of either sequence length, but not smaller than 30 amino acids.  The user can check which proteins to show in the structural alignment, and then obtain the alignment with EID sequences aligned to the PDB sequences. It should be noted that aligning the EID (GenBank) sequences does not change the structural alignment. A file is created in a format which allows Friend to display both the three-dimensional structures superimposed by the structural alignment in the structure window and the aligned sequences in the sequence window (A3). The exon boundaries and intron phases are denoted by color, red = phase 0, green = phase 1, blue = phase 2.
           The second main search is a sequence search which uses BLAST to find similar EID sequences. When performing this type of search a user can specify a PDB code, chain identifier, the number of output alignments, un/gapped BLAST, name of the database (experimental or non-experimental), and the number of HSPs (high-scoring segment pairs) to display.  Once a search is completed a webpage is displayed (B1) showing hits in a graphical output, along with a plot displaying the scoring information about the hits.  From this page a user can automatically visualize data in the Friend viewer similar to the output from the structural search. The sequence window (B2) is displayed in a multiple pair-wise alignment format, where all hits are aligned to the query sequence.

           Another feature of SEDB (query by Protein ID) provides users with the ability to map exon data for a protein sequence with unknown structure on similar proteins with known structures. This allows spatial visualization of exon/intron boundaries in a protein, without an experimental structure.