Plant, Animal & Microbe Genomes X Conference Plant, Animal & Microbe Genomes X ConferenceJanuary 12-16, 2002
Town & Country Convention Center
San Diego, CA
Bioinformatics: Algorithms
In day-to-day sequence analysis, biologists need to compare a new laboratory sequence of unknown function to an entire database of sequences, in order to find any related sequence. The speed of the sequence search depends on the size of the database. Sequence databases are often highly redundant for two basic reasons (1) due to an inefficient merge and storage of databases with identical and nearly identical sequences, and (2) often evolutionary mechanisms such as gene duplication and propagation are responsible for the existence of non-identical sequences with high sequence identity. Redundancy affects the sensitivity and speed of search algorithms and diminishes the benefit of searching large databases. The speed of sequence search can be improved by developing non-redundant sequence databases excluding identical and almost identical sequences. We report the construction of a simple non-redundant sequence database from various primary databases. The construcused database (RSDB.1) was based on parts of Swissport and Gene bank and and in conjunction with PSI-BLAST we used the Object oriented algorithm developed by us to remove neighboring sequences and redundancies. This report is an attempt to present a partial view on the relationship among size of the database, level of redundancy and overall sensitivity of sequence similarity search. Our work shows the benefit of employing constructed intermediate smaller RSDBs using appropriate search algorithms to conduct sequence similarity searches and retrieval of information from larger databases is more efficient with out having to bother of their size and level of redundancy.