We have created an algorithm for compressing a PIR database to assist individual researchers and software developers who utilize sequence database information but may not have huge storage space. The resulting compact databank contains compressed PIR information and an interface written in C which allows fast direct access to the stored information without extensive decompression of corresponding files. The databank files as well as the interface C-file can be used on both PC-compatibles and UNIX-based computers without any modifications. The interface supports all standard PIR Request Network queries (i.e. gets databank SEQ number by entry; for a defined databank SEQ number, gets specified information like: name, organism(s), keyword(s), sequence, sequence features with coordinates, etc.). In contrast with PIR Request Network, our package allows us to call PIR-contained information directly from the C programs, even on a personal computer not on a network. Our PIR-derived databank, SAGITTARIUS PIR, was implemented in the form of separate file sets. Each file set contains database information of independent types (i.e. sequences, entry indexes, organisms, etc.). On a particular computer, the available configuration of the PIR information (and storage space) can be easily changed as needed by the user without affecting retrievals of other types of stored information. Due to an original alignment-based algorithm, in the compression of protein sequences themselves, our package out-performs the well-known ZIP file compressor. For PC-compatibles, a dialogue shell is available which supports all standard PIR Request Network queries plus homology searches, alignments, etc.

A new algorithm for data bank homology search is proposed. The principal advantages of the new algorithm are: (i) linear computation complexity; (ii) low memory requirements; and (iii) high sensitivity to the presence of local region homology. The algorithm first calculates indicative matrices of k-tuple 'realization' in the query sequence and then searches for an appropriate number of matching k-tuples within a narrow range in database sequences. It does not require k-tuple coordinates tabulation and in-memory placement for database sequences. The algorithm is implemented in a program for execution on PC-compatible computers and tested on PIR and GenBank databases with good results. A few modifications designed to improve the selectivity are also discussed. As an application example, the search for homology of the mouse homeotic protein HOX 3.1 is given.

We present a new pairwise alignment algorithm that uses iterative statistical analysis of homologous subsequences. Apart from the classical conversion of the DOT-matrix characteristic of the Needleman-Wunsch algorithm (NW), we used only those matrix elements that corresponded to the most non-random subsequence homologies. The most reliable elements of the DOT-matrix are written to the compact competition matrices. The algorithm then searches for alignment on the base of only these matrix elements. Our algorithm has low storage and memory requirements, but provides a reliable alignment for the sequences of weak homology (or, at least for the homology regions). In such cases classical NW algorithms often produce unreliable results on the level of statistical noise due to accumulation of random matchings throughout the aligned sequences.