MOTIVATION: We propose representing amino acids by bit-patterns so they may be used in a filter algorithm for similarity searches over protein databases, to rapidly eliminate non-homologous regions of database sequences. The filter algorithm would be based on dynamic programming optimization. It would have the advantage over previous filter algorithms that its substitution scoring function distinguishes between conservative and non-conservative amino acid substitutions. RESULTS: Simulated annealing was used to search for the best five-bit or three-bit patterns to represent amino acids, where similar amino acids were given similar bit-patterns. The similarity between amino acids was estimated from the BLOSUM45 matrix. Representing amino acids by these five-bit and three-bit patterns, the Escherichia coli PhoE precursor and the bacteriophage PA2 LC precursor were aligned. The alignments were nearly the same as that obtained when BLOSUM45 was used to score substitutions. AVAILABILITY: The C code of the optimization algorithm for searching for the optimal bit-pattern representation of amino acids is available from the authors upon request.

A method exploiting quaternary integer arithmetic for the rapid comparison of nucleotide sequences is proposed. A central feature of the method is the identification of a suitable integer arithmetic operation which, operating on two numerical strings (representing two sequences), yields an output string unambiguously related to the number of matching nucleotides in the compared sequences. The number of matching nucleotide bases in compared sequences can be determined without having to make individual base by base comparisons. The method is a general extension of the algorithm previously described in the code UNIREP. The rules governing integer arithmetic as applied to nucleotide sequences are discussed.