SAGITTARIUS DNA Block Marker **************************** SAGITTARIUS DNA Block Marker is a package for exon/intron structure revealing on the base of protein k-tuples statistic, with orientation on MS DOS PC-compartibles (386 or 486 recommended). Input: any sequence file (.SEQ). Probably (not well tested..) you can use free format sequence representation, GCG-style files, standard sequence extractions from GENBANK or EMBL databases etc. Output: on-screen and in file EXONS.RES. Output includes probable coding regions list (start-end, frame#, weight) with corresponding aminoacid sequences (this is special feature of this method - even in the case of nonperfect region boundary prediction error in coding frame is practically impossible), and suggested variant of assembled gene sequence (start-end and frame# for all included regions) with full coded aminoacid sequence. Alternative input: in presense of SAGITTARIUS GENBANK on your computer you can use on input any file with bank numbers (.NOM, standard output of buffer content from SAGITTARIUS databanks). Program will predict coding regions and genes for corresponding database sequences in parallel with demonstration of real GENBANK CDS features for each sequence what allows to test prediction quality on the set of well-known genes (learning mode). In the case of user sequence input (from file .SEQ) program will search for coding regions in both of direct and inverted strands consequently (independent predictions). ------------------------------------------------------------- Short algorithm description --------------------------- Taking into accout a certain proximity of the k-tuple (peptide) organization of proteins with similar structure and/or functions, a new method is proposed for detection of coding regions of genome based on mapping of newly sequenced DNA regions by subsequences "admissibility" function. This function obtained by k-tuple analysis of protein data base. In the simplest case, this algorithm marks all the k-tuples not found in real proteins as structural prohibitions which are forbidden in the coding regions of analogous proteins. The longest regions without such prohibitions can be realized as the best exon (for eukaryotic genome) or active gene (for prokaryotic genome). To ensure a statistically significant level in the comparison of different sequences, the length k of the used k-tuples is chosen taking into account the general incidence of respective peptides in the data base. Thus, for chance sequences with the same amino acid composition as the data base under study, all the used k-tuples must be realized in the data base at least once. In this case, the absence of some k-tuples can detect both the limitations on the primary protein structure and a mere lack of statistical data. The proposed algorithm has been tested on the basis of the latest releases of data bases EMBL and SwissProt. As the best examples, human genes with a complicated exon/intron structures were tested. Practically in all cases, the longest region without any k-tuple prohibitions was present in the database EMBL as CDS feature. For all the other shorter real exons, the same result was observed, although with a lower signal/noise ratio. The position forecast for the longest exon was much higher than in the case of using other coding potentials (up to 95% instead of 60-70%). The field of application of the proposed algorithm is somewhat limited by the existence of a structural or functional analog of the coded protein among the already sequenced proteins in the data base. The probability of such a failure has been estimated by means of checking the number of appearing new superfamily proteins in two consecutive releases of PIR base and found to be 2-5%. In all other cases the algorithm may be used as a tool for a rapid and accurate detection of at least one best exon in a eukaryotic genome (with subsequent more detailed study of complete exon structure by means of usual algorithms) or a complete active gene in a prokaryotic one. ------------------------------------------------------------- SAGITTARIUS DNA Block Marker is available by anonymous FTP from: - FTP.SCRI.FSU.EDU, directory /pub/genetics/exons/ Probably SAGITTARIUS DNA Block Marker is available by anonymous FTP from some of the well-known bio-servers (iubio etc.). ------------------------------------------------------------- Installation ------------ Distributive variant includes ready-for-use informational files and executables (2.5 Mb in total) - all in self-extracting archieve (file exonsall.exe). All decompressed SAGITTARIUS DNA Block Marker files must be placed in one directory (including Borland .BGI and huge .WRK files). Program intensively read data from huge file .WRK what slow down prediction process. If you have on your computer virtual (in-memory) disk of at least 2.1 Mb capasity you may copy huge file .WRK on this disk before program running and assign this disk drive letter G:. Such operation allows program to find .WRK file copy on this (G:) virtual disk and speed up process at least 5 times. In absense of such large virtual disk program will use .WRK file from current directory. -------------------------------------------------------------- SAGITTARIUS is a FREE DOMAIN software. This package (with compressed data files) can be redistributed freely without any limitations but only free of charge and for non-commercial usage. No changes in data files and/or executables allowed. -------------------------------------------------------------- For HELPFUL comments and discussions please contact: Dr. Victor B.Strelets (strelets@scri.fsu.edu) or Dr. Hwa A.Lim (hlim@scri.fsu.edu) Supercomputer Computations Research Institute, Florida State University, B-186, Tallahassee, FL 32306-4052, USA --------------------------------------------------------------- Common SAGITTARIUS information ------------------------------ SAGITTARIUS is a family of free domain application packages for molecular biologists with orientation on MS DOS PC-compartibles: - Compressed sequence databases with dialog shells for fast and easy data manipulation (PIR and GENBANK variants) - Fast programs for cross-bank user SEQ homology searches (short subregion homologies sensitive) (closely connected with compressed SAGITTARIUS databases) - Fast sensitive programs for pairwaise alignments (both aminoacid and nucleotide), including cross-bank user SEQ alignments (closely connected with compressed databases to allow cross-bank user SEQ alignments) - Packages for fast tree-based multiple alignments - Package for sequenation-errors-stable contigs joining (on the base of tree-based multiple alignment) - Automated system for revealing coding regions (exon/intron structure) in new nucleotide sequences (including learning mode access to the nucleotide sequences in compressed SAGITTARIUS GENBANK database) - Personal Reference Database dialog shell for manipulation of data from standard BIO-JOURNALS(BIOSCI), SEQANALREF(Bairoch), JOURNALS-TOC(multiple sources) databases -------------------------------------------------------------- Author(s) will in no way be held liable for any loss of profit or any other commercial damage including but not limited to special, incidental, consequential or other damages from use of this package. You may use them only with the understanding that you use it at your own risk and that your use of the software is your agreement to this disclaimer.