Rationale

"Science is built of facts the way a house is built of bricks; but an accumulation of facts is no more science than a pile of bricks is a house." --Henri Poincaré "Studies have suggested that the regulation of genes rather than the genes themselves is responsible for the differences between humans and their closest relatives." --from Chimp Genome Says Much about Human Evolution "Thales of Miletus, creator of science in Western civilization... responded to a challenge to show that science is of value in practical business matters." --from History of the Ionian Civilization "A philosopher who cannot grasp and command actuality as well will never be of the first rank." --Oswald Spengler "Get out of your rut, open your mind - there's a whole Universe waiting. It's waiting for people who aren't afraid of thinking in new ways, of doing new things." --Isaac Asimov

In the beginning was the vision...

"A researcher in cell signaling wakes up in the morning to attend a meeting where the most recent findings in her field will be presented. As slides change rapidly, she uses a special symbolic language to write down all the new information. Better yet, she directly types it in a text document opened on her laptop... In the afternoon, she finds time to read a research article. There is a lot of new information on her favorite signaling pathway in this article. As she reads on, she appends the new data to the document she has created on her computer at the meeting -- again, in this simple, intuitive signaling pathway language. She is now just one click away from displaying a graphical map of the signaling molecules and interactions in the pathway: from any Web-enabled computer she can log in to a program at a Web site that allows her to copy-paste her data into a browser window to display and save the graphical map. And tomorrow, or at any time when more data is available, to update the signaling pathway map all she will need to do is add a few more lines to her text document, then... click! again. Now imagine the map can be edited and customized in the vector graphics presentation software installed on her laptop computer..."

This is now possible.

The Dynamic Signaling Maps™ informatics biotechnology system provides the cell signaling community with a new and powerful means to encode and store experimental data, check for and eliminate data inconsistency, and graphically render the signaling data as maps, or diagrams, which can be further edited and customized.

Three major trends have impinged on the development of these tools.

First, the hallmark of the postgenomic era in biotechnology is the rapid pace of research in functional genomics and proteomics. The simplistic definition of the "proteome" refers to the total set of protein species present in a biological cell. A more complete definition also includes the functional component consisting of the interactions among these proteins; these interactions are the bedrock of the physiological functions of any cell and are globally referred to as the "physiome". A list of all proteins and interactions, for any particular species or cell type, is still far from completion. While several projects are under way to map these multiple interactions, the need exists for methods and means to visualize them in a simple and intuitive manner. Given the amount of data, the classic approach consisting in the manual drawing of schematics using presentation software on desktop computers is no longer appropriate.

Second, experimental data in biomedicine is often factual, enumerative, redundant, "patchy", sometimes even contradictory, and virtually always expressed in common rather than formal language. Biological signaling data are no exception to this rule (e.g., "A activates B.") and a number of natural language processing tools exist that can convert literature data into machine-readable formats with various degrees of accuracy. Data extracted using these tools still need curation by qualified human operators and are not free of logical inconsistencies. This is why the signaling community would become more efficient if a means existed for stepping up the rate of detecting and eliminating data inconsistencies and redundancy. The need also exists for a tool that could update complex signaling maps, especially as new experimental data becomes available and previously valid signaling maps become obsolete.

Third, the characteristics of experimental data mentioned above make formal, operational models of cellular functions hard to come by. Access to an informatics biotechnology system that can produce a consistent, clear map of signaling data would foster the development of comprehensive formal models of cell signaling and protein networks.

A number of public biomolecular interaction repositories exist: BIND; MINT; IntAct; DIP; HPRD; subsets of KEGG and CSNDB. These databases contain information on the basic building blocks of biological signaling pathways. Based on this information, molecular interaction clusters can be generated; they are also referred to as "biological association networks" in some circles. But a molecular interaction cluster does not represent a signaling pathway per se. In effect, more information about each interaction, such as its outcome (e.g., activation, inhibition), is required for it to become a bona fide component of a signaling pathway. Also, equal importance is given in an interaction database to a complex and to a binary interaction, while in a signaling pathway, the complex usually plays a role similar to a single molecule and the binary interaction often represents a signaling event. Moreover, other fundamental components of signaling pathways, such as translocations and reactions, are generally omitted from biomolecular interaction databases, which makes perfect sense. For all purposes and means, signaling pathways represent a level of organization and knowledge above mere biomolecular interactions, therefore they require appropriate tools for encoding, visualization, and analysis. We have addressed this unmet need of the Life Science benchwork scientist in early 2002 by developing the Dynamic Signaling Maps™ Web-based software suite.

Dynamic Signaling Maps™ comprises the following components and functions :

1. Pathway Editor
1. enter signaling pathway data using a simple symbolic language (the "DSM Language")
2. store/retrieve information to/from user files on any workstation with Web access
2. DSM Language Translator
1. check for syntax errors and provide efficient editing means
2. detect and eliminate data inconsistencies (e.g.: contradictory, or incompatible, data)
3. detect and eliminate data redundancies
4. detect similarities and issue warnings
5. provide users with "clean" versions of their data in a format suitable for database storage
6. support pathway merging
3. Pathway Mapper
1. render output from Translator in editable vector graphics formats

Modules :

4. Signaling Event Analysis
1. analyze pathways downstream and upstream of given discrete signaling events
2. integrate quantitative simulation capabilities where possible (in development)
5. Molecular Interaction Clusters Search & Visualization
1. search, visualize molecular interaction networks compiled from curated public databases
2. provide links from pathways map objects to a comprehensive literature/reference database
6. Gene Network Extractor
1. compute gene networks from time-series expression profiles
2. dynamically overlay expression profiles onto gene network diagrams
7. Signaling Pathway/Gene Network Manager
1. extend items 2-6 above to multiple, translated pathways in the context of a local database (in development)

•  Signaling pathways and biochemical pathways are different categories, or object classes. Comparing signaling/regulatory pathways with metabolic/biochemical pathways would be similar to comparing the military chain of command to the chain of supply. It is irrelevant how many generals decide to carry out a particular battle and give the order that sets the troops in motion; what matters is that the order be received in the trenches and executed. Likewise, when analyzing biological signaling pathways, it is irrelevant how many identical molecules do trigger a signaling event; what matters is that the signaling event occur, that is the relevant biological event.
•  However elegantly designed and intelectually stimulating, numerical simulation methods in Biology have been so far virtually useless in practice due to the sheer complexity of the systems involved and the number of parameters to know for the mathematical simulators to work. Consider the relatively simple Belousov-Zhabotinsky (BZ) reaction, a biological system-like chemical oscillator that has been mathematically modelled by several groups over several decades; actual biological systems are several orders of magnitude more complex in terms of molecular species and interactions in any cell compartment. Biomedical research scientists and physicians are intuitively reluctant to consider using numerical approaches in their investigations unless the methodology in cause is clearly validated in practice.
•  Biological systems are heterogenous and discontinuous at the molecular level. Kinetic models of biochemical reactions deal with continuous mathematical variables and therefore have limited applicability to biological systems in which only one or few molecules are involved. A relevant example is immunoglobulin gene rearrangement, where the enzymatic modification performed by a single RAG recombinase complex, at a single gene locus, on one chromosome only, decides the fate of the B lymphocyte undergoing that particular developmental process. To our best knowledge, no type of kinetic model that could simulate such a discrete event exists; however, this theoretical limitation may be overcome in the future by new advancements in Mathematics.
•  There are already at least two XML-based simulation and data exchange languages for biochemical reactions and pathways, SBML and CellML, which were designed to be primarily machine-readable (i.e., parsable using computer programming languages). Importantly, the DSM Language is both human-readable and machine-readable; this means a DSML document can be, on the one hand, easily read-interpreted-edited by the investigator and, on the other hand, automatically parsed for analysis and visualization with a machine. While the Dynamic Signaling Maps tools can generate pathway diagrams in the XML-based format SVG, an export-to-SBML function for pathways is in development to facilitate interfacing with third party utilities.

Copyright © 2002-2010 Hippron Physiomics Inc. All rights reserved.