Schematic diagram of construction of the PEDE database. Tables composing the database that are used for the web interface are listed. The data type of each column is indicated in PostgreSQL format (text, integer, char and real). Columns linked with lines are dependent on each other. An estimate of whether contigs (or singlets) include full-length CDSs and the detected SNPs alter coding amino acids can be inferred by using the values in the tables, which are available on demand. UniGene, RefSeq, and LocusLink data were obtained from NCBI (8) and integrated into the database.