Motivation: A rapid discovering of experimental data on regulation of gene expression and their accumulation in various databases accompanied by development of numerous and diverse software tools for their analyses demands integration of the available informational and software resources.

Introduction

Currently, the number of databases on gene expression and a variety of software for the analysis of these data are growing fast. An Internet-accessible system GeneExpress-2.1 is being developed for accumulation of experimental data, their analysis, and navigation through integrated software and informational resources related to regulation of gene expression. It integrates a large amount of databases and hundreds of programs for processing the data on the structure-function organization of DNA, RNA, proteins, and gene networks together with other informational resources important for Informational resources of GeneExpress.

The structure of the GeneExpress-2.1 corresponds to the natural hierarchical organization of molecular genetic systems, containing the following levels: (1) DNA level, (2) RNA level, (3) protein level, and (4) gene network level. Each module contains (1) experimental data represented as a database or a sample; (2) program for data analysis; (3) results of an automated data processing; and (4) tools for graphical representation of these data and results of the data analyses. The main databases of GeneExpress-2.1 involve the relational model of data representation. For access to the databases of GeneExpress-2.1 system, RDBMS (Relation Database Management System), ORACLE 9i, and Sequence Retrieval System (SRS 6.0) are used.

Research into mechanisms underlying molecular interactions depending on the genetic information and specific features of molecular structures may shed the light on biochemical functions and roles of elementary components as well as on the specific control patterns of gene networks. These pieces of knowledge form the background for computer simulation of gene networks allowing changes in molecular genetic, biochemical, physiological, morphological, and other characteristics of various organisms to be predicted as well as optimal control actions and stimuli for correcting genetically specified impairments of the body operation to be searched for.

For this purpose, a new generation computer technologies integrated in the computer system GeneNetDiscovery is developed at the Siberian Branch of the Russian Academy of Sciences. This system provides solving a wide range of problems in the field of computer analysis and simulation of complex molecular genetic systems (gene networks, genetically controlled metabolic pathways, signal transduction pathways, etc.) including (i) accumulation of data and knowledge on the structure-function organization of gene networks; (ii) integration of the information on gene networks and metabolic pathways; (iii) construction of gene network mathematical models and their computer-assisted numerical analysis; (iv) study of dynamic behavior of complex molecular genetic systems (gene networks) in norm, in case of pathologies and metabolic diseases, and under the effect of adverse environmental factors at molecular genetic, cellular, and organismal levels; and (v) search for optimal control of gene networks and correction of their behavior in the case of various pathological states.

1.      Transcription Regulatory Regions Database (TRRD)

TRRD is designed for accumulation of experimental information on the structure-function organization of regulatory regions of eukaryotic genes. It is a unique informational resource on long gene transcription regulatory regions. In addition to description of regulatory region itself, it provides (a) description of the hierarchy of all the regulatory units included into a described regulatory region (such as transcription factor binding sites, promoters, enhancers, silencers, etc.); (b) information on expression patterns of the genes described; and (c) information on physiological systems, organs, and cell types wherein these genes are expressed. The new release of TRRD-6.0 contains interferon-inducible genes; erythroid-specific genes; genes of lipid metabolism in liver, adipose tissue, at the cell and organismal levels (cholesterol regulation, and leptin hormone regulation, lipid exchange between lipoprotein blood particles); glucocorticoid-inducible gene;, cell cycle-dependent genes; genes of the endocrine system; heat shock-regulated genes; redox sensitive genes; iron metabolism genes; macrophage-expressed genes;  apoptosis genes; and plant genes. TRRD-6.0 comprises descriptions of 1405 genes, 6646 sites, and 2158 regulatory regions. This database is supported constructed using ORACLE 9i, and Sequence Retrieval System (SRS 6.0) is used to access TRRD-6.0. A novel version of the TRRD Viewer (release 2.0) implemented as a Java-applet allows the regulatory gene regions described in TRRD to be visualized.gene expression regulation. From its first version (Kolchanov at la., 1998a; 1998b), GeneExpress is intensively developing. This paper briefs the state of GeneExpress –2.1 in 2002. Descriptions of its individual modules in more detail are available in papers included in Proceedings of BGRS’2002. The system is widely and actively used for computer analyses of various organizational levels of molecular genetic systems.

2.      Programs for recognizing regulatory elements involved in controlling the transcription

GeneExpress-2.1 has a large set of original programs for recognition and various analyses of transcription factor binding sites and promoters as well as study of specific contextual and structural DNA features in gene regulatory regions, exemplified below.

Resource	Description
BinomSiteMMSite	Searching for potential transcription factor binding sites (TFBS) using a binomial criterion for estimating similarity scores between regions of a sequence analyzed and the TFBS sequences described in TRRD URL: http://wwwmgs.bionet.nsc.ru/mgs/programs/mmsite/http://wwwmgs.bionet.nsc.ru/mgs/programs/multalig/
MultipleRecognition MMSite (MultRec)	Simultaneous usage of the entire set of recognition methods for detecting TFBSs. URL: http://wwwmgs.bionet.nsc.ru/mgs/programs/multalig/
ARGO_ViewerARGO_Reg_Finder	Recognition of promoters of tissue-specific gene groups basing on the analysis of the presence of specific quasi-invariant oligonucleotide motifs detected using the program ARGO. URL: http://wwwmgs.bionet.nsc.ru/mgs/programs/argo/argo_viewer.htmlSARs/MARs recognition based on the specific quasi-invariant oligonucleotide motifs. URL : http://wwwmgs.bionet.nsc.ru/mgs/programs/argo/argo_viewer.html
RGSiteScanARGO_Nucleo_Finder	Searching for TFBSs basing on the recognition group approach. URL: http://wwwmgs.bionet.nsc.ru/mgs/programs/yura/RecGropScanStart.htmlSearching for potential nucleosome formation sites basing on the presence and relative location of specific quasi-invariant  oligonucleotide motifs detected by the program ARGO. URL : http://wwwmgs.bionet.nsc.ru/mgs/programs/argo/argo_viewer.html
KD_PromSite_Annotator.	Recognition of RNA PolII promoters from their contextual patterns determined using knowledge discovery and data mining in TRRD. URL: http://wwwmgs.bionet.nsc.ru/mgs/programs/recon2/Simultaneous recognition of various types of the TFBSs in an unknown sequence using all the above-described programs (MultRec, RGSiteScan, MMSite, LSsite, ARGO_Site_Finder). URL :
ProGAKD_Prom	Recognition of RNA PolII promoters. URL: http://wwwmgs.bionet.nsc.ru/mgs/programs/proga/Recognition of RNA POLII promoters from their contextual patterns determined using knowledge discovery and data mining in the TRRD database. URL :
BLAST_PromoterProGA	Recognition of the RNA PolII promoters basing on the BLAST search for homology with the promoters described in the TRRD. URL: http://wwwmgs.bionet.nsc.ru/mgs/systems/fastprot/blast.htmlRecognition of RNA POLII promoters. URL :http://wwwmgs.bionet.nsc.ru/mgs/programs/proga/
ReconBLAST_Promoter	Searching for potential nucleosome formation sites basing on the nonuniformity of dinucleotide context within local promoter regions. URL: http://wwwmgs.bionet.nsc.ru/mgs/programs/recon/Recognition of the RNA POLII promoters basing on the Blast-search for homology with the promoters described in the TRRD URL :http://wwwmgs.bionet.nsc.ru/mgs/systems/fastprot/blast.html

3.      Informational and software resources on RNA structure-function organization

To solve a variety of problems on analyzing RNA structure-function organization, a number of databases and software tools were developed with GeneExpress-2.1 and united in the module RNA Integration Level. It comprises (i) a number of programs for calculation of RNA secondary structure and evaluation of the secondary structure formation potential and (ii) the knowledge base on structure-function organization of leader mRNA sequences.

Resource	Description
Program GArna	Applying genetic algorithm to predict the secondary structures displaying seals energies and visualize them. URL: http://wwwmgs.bionet.nsc.ru/mgs/programs/2dstructrna/
Program MatrixSS	Calculation of E score, a contextual characteristics reflecting the potential for forming RNA secondary structure compared with random sequences. URL: http://wwwmgs.bionet.nsc.ru/mgs/programs/2dstructrna/MatrixSS.html
Knowledge base LEADER_RNA	LEADER_RNA is a tool to evaluate mRNA translational properties. Contains a database with samples of 5'UTR sequences of high- and low-expressed mRNAs of mammals, dicot, and monocot plants. These sequences are used as training samples for the computer system. This knowledge base contains also (1) description of the discovered mRNA properties that may be used to discriminate between the high- and low-expressed mRNAs and (2) programs predicting mRNA translational efficiencies from significant contextual and structural characteristics of mRNA 5'UTRs (C codes for prediction of mRNA translation level). URL: http://wwwmgs.bionet.nsc.ru/mgs/gnw/leader/

4.      Informational and software resources on the structure-function organization of proteins

A number of databases and software tools, forming the modules Protein Integration Level of GeneExpress-2.1, have been developed for solving the problems related to analyses, structure, function, and evolution of proteins. This module contains the databases on (i) expanded annotation of the EnPDB-compiled structures, (ii) active sites of the PDBSite-compiled proteins, (iii) protein and peptide sequences obtained by artificial in vitro selection (ASPD) as well as the programs for (iv) searching the protein spatial structure for re regions similar to PDBSiteScan-compiled active centers and (ii) detecting and analyzing the coordinatively fixed amino acid substitutions (CRASP, Gene Network Level)

Resource	Description
Database EnPDB	Expanded options for indexed search for information in the PDB databank entries URL: http://wwwmgs.bionet.nsc.ru/mgs/gnw/enpdb/
Database PDBSite	Information on the spatial structure and physicochemical properties of 4723 active protein sites annotated in PDB. URL: http://wwwmgs.bionet.nsc.ru/mgs/gnw/pdbsite/
Program PDBSiteScan	Searching for active sites in protein spatial structures according to the sequence and spatial arrangement of amino acid residues. URL: http://wwwmgs.bionet.nsc.ru/mgs/systems/fastprot/pdbsitescan.html
Database ASPD	Information on peptide and protein sequences produced an in vitro selection. URL: http://wwwmgs.bionet.nsc.ru/mgs/gnw/aspd/
Program CRASP	Detection of coordinated amino acid substitutions in protein families and analysis of their physicochemical characteristics. URL: http://wwwmgs.bionet.nsc.ru/mgs/gnw/crasp/

5.      Informational and software resources on the structure-function organization and operation dynamics of gene networks

This module of GeneExpress-2.1, GeneNetDiscovery, is a tool designed for creating gene network models that could be further used in another operation system and for a variety of purposes.

For annotators and modelers, specialized worksites are developed in a Windows NT/2000/XP environment. Web interfaces are developed for outside “casual” users. Currently, two client programs are realized (gene network editor and gene network viewer), which communicate with the server by HTTP protocol realized as exchange by the XML messages.over HTTP protocol, by which XML messages are sent. Oracle9i is appled for controlling the databases. When developing the system’s middleware controlling the logics of its operation and its linkage to the knowledge bases and databases, application server Oracle9iAS is used; it includes Container for J2EE (OC4J), XML SQL Utility, and XML Parser as components. The operation logics of the system GeneNetDiscovery is shown in Fig. 1.

Fig. 1. Functional layout of the system GeneNetDiscovery.

The system GeneNetDiscovery comprises the following functional modules:

(1)   Subsystem for constructing models, including

• GeneNet database, which accumulates the experimentally confirmed information on the structures of gene networks, compartmentalization of their components at the levels of cells and overall organism, functional interactions of gene network components, signal transduction pathways, and metabolic pathways, dynamical data, etc.. (Ananko et al., 2002). URL: http://wwwmgs.bionet.nsc.ru/mgs/gnw/genenet/GeneNet-0002.shtml

A new version of this database in an Oracle9i environment is under development now (Loktev et al., 2002);

• Module GENED (gene network editor) for visual simulation of the structure-function organization of complex gene networks and data input into GeneNet (Loktev et al., 2002); and
• Module for interactive construction of gene network models using the information accumulated in the GeneNet database (as systems of differential equations, logic automatons, or hybrid models describing gene network dynamics);

(2)   Subsystem for analyzing gene network models, including

• Module for computer logical analysis of the structure-function organization of gene network basing on a graph-theoretic approach (Dobrynin et al., 2002) and
• Module GeneNetStep for analyzing numerically gene network models and studying qualitatively the behavior of solutions of differential equations in the space of model’s parameters (Fadeev et al., 2002);

(3)   Subsystem for identifying models, including

• Module for solving inverse problems, that is, identifying models and assessing their parameters from experimental data, expert estimations, and experimentally observed trajectories of system behaviors (Likhoshvai et al., 2002); and
• Database of dynamical data describing behavior of molecular genetic subsystems under various experimental conditions (Likhoshvai et al., 2002);

(4)   Subsystem for simulating gene networks and analyzing their behavior, including

• Module for calculating trajectories of behavior of molecular genetic systems and accumulating the calculated variants in database. Simulation of gene network dynamics; analysis of gene network operation when concentrations of individual components or rates of individual reactions are changed; simulation of effects of additional components on the gene network operation; and simulation of mutations.

URL: http://wwwmgs.bionet.nsc.ru/mgs/gnw/gn_model/modelling.shtml

• Module for computer analysis of gene network dynamics basing on data mining and knowledge discovery approach; search for gene network behavior patterns, their analysis, and generalization for creating the knowledge base with the system (for testing of certain algorithms see Borisova et al., 2002);
• Knowledge base compiling variants of the calculated trajectories of gene network behaviors and pieces of knowledge used for constructing models;

Sets of differential equations describing dynamics of reactions, reaction rate constants, and initial concentrations of the components for dynamical models of three gene networks, namely, gene network regulating (1) lipid metabolism, (2) erythrocyte differentiation and maturation under the effect of EPO, and (3) activation of macrophages by IFN-g and LPS.

URL: http://wwwmgs.bionet.nsc.ru/mgs/gnw/gn_model/

• Application of the models developed for solving various scientific problems, in particular, studying the effect of mutations on gene network dynamics and molecular mechanisms underlying various pathologies stemming from impairments of the gene network operation (Ananko et al., 2002; Stepanenko and Grigor’ev, 2002; Turnaev and Podkolodnaya, 2001; Nedosekina and Ananko, 2002; Kudryavtseva and Stepanenko, 2002; and other); and

(5)   Subsystem for controlling gene networks, including

• Module for analysis of mutational portraits of gene networks and detection of optimal targets for their pharmacological regulation (Ratushny et al., 2002) and
• Module searching for optimal control of gene network operation in norm and pathology, in particular, for correction of organismal pathological states with account of their individual genotype-specific features (Latypov et al., 2001a; 2002b)

Conclusion

Described in the work is the architecture of the system that is developed basing on multilevel computational approaches combining genome-encoded information with nongenomic network connections. A part of the modules of the system GeneNetDiscovery has been realized. The major algorithms forming the core of this system have been tested by solving particular problems.

Acknowledgements

Work was supported in part by the Russian Foundation for Basic Research (grants No. 01-07-90376, 01-07-90084, 00-07-90337, 02-07-90355, 02-07-90359, 00-04-49229, and 00-04-49255); Russian Ministry of Industry, Science, and Technologies (grant No. 43.073.1.1.1501); Siberian Branch of the Russian Academy of Sciences (Integration Projects Nos. 65 and 66) US National Institutes of Health (grant No. 2 R01-HG-01539-04A2); and US Department of Energy (grant No. 535228 CFDA 81.049).

References

1.      Kolchanov, N.A., Ponomarenko, M.P., Kel, A.E., Kondrakhin, Yu.V., Frolov, A.S., Kolpakov, F.A., Kel, O.V., Ananko, E.A., Ignatieva, E.V., Podkolodnaya, O.A., Stepanenko, I.L., Merkulova, T.I., Babenko, V.N., Vorobiev, D.G, Lavryushev, S.V., Ponomarenko, Yu.V., Kochetov, A.V., Kolesov, G.B., Podkolodny, N.L., Milanesi, L., Wingender, E., Heinemeyer, T., and Solovyev, V.V. (1998a). GeneExpress: a computer system for description, analysis, and recognition of regulatory sequences of the eukaryotic genome. ISMB,. 6:95-104. MEDLINE PMID: 9783214; UI: 98456543.

2.      Kolchanov, N.A., Ponomarenko, M.P., Kondrakhin, Yu.V., Frolov, A.S., Kolpakov, F.A., Kel, A.E., Kel-Margoulis, O.V., Ananko, E.A., Ignatieva, E.V., Podkolodnaya, O.A., Stepanenko, I.L., Merkulova, T.I., Babenko, V.N., Vorobiev, D.G., Lavryushev, S.V., Grigorovich, D.A., Ponomarenko, J.V., Kochetov, A.V., Kolesov, G.B., Podkolodny, N.L., Wingender, E., Heinemeyer, T., Milanesi, L., Solovyev, V.V., and Overton, O.K. (1998b). GeneExpress system: description, analysis, and recognition of regulatory sequences in eukaryotic genomes. Proc. I Intern. Conference on Bioinformatics of Genome Regulation and Structure, BGRS’98, Novosibirsk–Altai Mountains, Russia, August 24-31, 1998, 71-76.

3.       Afonnikov, D.A. (2002). Contribution of coadaptive substitutions to the stability of physicochemical properties of ATP-binding sites in protein kinases. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

4.      Afonnikov, D.A., Oshchepkov, D.Yu., and Kolchanov, N.A. (2001). Detection of conserved physico-chemical characteristics of proteins by analyzing clusters of positions with coordinated substitutions. Bioinformatics, 17, 1035-46.

5.      Ananko, E.A., Podkolodny, N.L., Ignatieva, E.V., Podkolodnaya, O.A., Stepanenko, I.L., and Kolchanov, N.A. (2002). GeneNet system: its status in 2002. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

6.      Dobrynin, A.A., Makarov, L.I., and Podkolodny, N.L (2002). A graph-theoretic approach to computer analysis of gene network structure. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

7.      Loktev, K.A., Tkachev, Yu.A., Ananko, E.A., and Podkolodny N.L. (2002). A system for visual modeling of gene networks structural and functional. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

8.      Likhoshvai, V.A., Latypov, A.F., Nedosekina, E. A., Ratushny, A.V., and Podkolodny, N.L. (2002). Technology of using experimental data for verification of models of gene network operation dynamics. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

9.      Borisova, I.A., Zagoruiko, N.G., Likhoshvai, V.A., Ratushny, A.V., and Kolchanov, N.A. (2002). Diagnostics of mutations based on analysis of gene networks. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

10.  Fadeev, S.I., Berezin, A.Yu., Gainova, I.A., Kogai, V.V., Ratushny, A.V., and Likhoshvai, V.A. (2002). Development of the program software for mathematic modeling of the gene network dynamics . Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

11.  Kudryavtseva, A.N. and Stepanenko, I.L. (2002). Gene network of glutathione homeostasis: a response to oxidation stress. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

12.  Latypov, A.F., Nikulichev, Yu.V., Likhoshvai, V.A., Ratushny, A.V., Matushkin, Yu.G., and Kolchanov N.A. (2002a). A method of solving problems of optimal control in dynamics of gene networks. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

13.  Latypov, A.F., Nikulichev, Yu.V., Likhoshvai, V.A., Ratushny, A.V., Matushkin, Yu.G., and Kolchanov N.A. (2002b). Problems of control of gene networks in a space of stable states. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

14.  Nedosekina, E.A. and Ananko, E.A. (2002). Gene network of macrophage activation under the action of interferon-gamma and lipopolysaccarides. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

15.  Ratushny, A.V. and Likhoshvai, V.A. (2002). Computer analysis of the effects of mutations in LDL receptor gene on the regulation of cholesterol biosynthesis in the cell. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

16.  Ratushny, A.V., Likhoshvai, V.A., and Kolchanov, N.A. (2002). Analysis of mutational portraits of gene networks. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

17.  Turnaev, I.I. and Podkolodnaya, O.A. (2002). Gene network on cell cycle control. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).

18.  Stepanenko, I.L. and Grigor’ev, S.A. (2002). Organization of the gene network of apoptosis. Proc. III Intern. Conference on Bioinformatics of Genome Regulation and Structure (BGRS'2002).