©
1996 Oxford University Press
26-29
Footnote
The metabolic pathway collection from EMP: the enzymes and metabolic pathways
database
The metabolic pathway collection from EMP: the enzymes and metabolic pathways database
Evgeni
Selkov
,
Svetlana
Basmanova
,
Terry
Gaasterland
1
,
Igor
Goryanin
,
Yuri
Gretchkin
,
Natalia
Maltsev
1
,
Valeri
Nenashev
,
Ross
Overbeek
1,
*
,
Elena
Panyushkina
,
Lyudmila
Pronevitch
,
Evgeni
Selkov, Jr
and
Ilya
Yunus
Institute of Theoretical and Experimental Biophysics, Russian Academy of
Sciences, 142292
Pushchino
,
Russia
and
1
Mathematics and Computer Science Division, Argonne National Laboratory,
Argonne
, IL 60439,
USA
Received September 26, 1995
;
Revised and Accepted October 26, 1995
ABSTRACT
The Enzymes and Metabolic Pathways database (EMP) is an encoding of the contents
of over 10 000 original publications on the topics of enzymology and
metabolism. This large body of information has been transformed into a
queryable database. An extraction of over 1800 pictorial representations of
metabolic pathways from this collection is freely available on the World Wide
Web. We believe that this collection will play an important role in the
interpretation of genetic sequence data, as well as offering a meaningful
framework for the integration of many other forms of biological data.
INTRODUCTION
The Enzymes and Metabolic Pathways database (EMP) is a database on the
biochemistry of some 1800 different organisms (
1
). Over 10 000 journal articles have been encoded into quantitative data for
almost all documented enzymes. The curation of the encoded data and of the
pictorial representations of pathways is an ongoing project centered at the
Laboratory of Mathematical Simulation of Multienzyme Systems at the Institute
of Theoretical and Experimental Biophysics of the Russian Academy of Sciences,
in Pushchino, Russia.
The effort to build EMP was initiated in 1984. The initial motivation was to
support internal projects in the mathematical simulation of cell metabolism.
The goal of the effort was to encode as much of the known data relating to
enzymology as possible. An attempt was made to define a database format that
allowed one to encode the types of factual assertions made by authors of papers
in enzymology (
2
). Once an initial format was defined, the process of encoding research papers
began.
When metabolic pathways were presented in papers, they were encoded into EMP.
Often, it was necessary to add the EC (enzyme nomenclature) numbers. As the
collection of pathways grew, it became possible for an increasing number of
cross-checks to be performed, and numerous variations of common pathways were
recognized. Each pathway has been labeled with EC numbers when known. In 1995,
these pathways were extracted from EMP and made freely available to other
researchers (see below for information on distribution). While this paper
describes the collection of over 1800 metabolic pathway diagrams, these
represent a small portion of the overall contents of EMP; the metabolic
pathways do indeed offer a useful functional overview, but the vast majority of
the content of EMP is detailed data relating to specific enzymes (encoded into
slightly over 300 distinct field types).
Contents of the Metabolic Pathway Collection
Currently, the metabolic pathway collection available to users via the World
Wide Web contains 1814 distinct pathway drawings. Each chart is associated with
one or more taxonomic groups. Table
1
gives the most represented taxonomic groups, along with the number of
associated diagrams.
There are over 250 other taxonomic groups for which fewer than 10 metabolic
pathways are included in the collection.
Table
1
shows the number of enzymes covered by the pathways for the three most
represented organisms (the reader should note that this number does not
represent the number of sequenced enzymes, but rather the number of distinct EC
numbers that occur in the metabolic pathway diagrams). The reader should be
aware that the same enzyme may appear in a number of charts.
We have made the pathway collection available in two forms:
1. It can be browsed in the context of the World Wide Web application PUMA,
which can be reached via the following URL:
http://www.mcs.anl.gov/home/compbio/PUMA/Production/puma.html
2. It can be acquired via anonymous FTP from the BioBase server (details below).
The collection available from the FTP site is broken into the following broad
categories for ease of distribution:
Amino Acid Metabolism
Aromatic Hydrocarbons
Carbohydrate Metabolism
Coenzymes and Vitamins
Electron Transport
Enzyme Metabolism
Halide Metabolism
Hydrogen Metabolism
Intracellular Transport
Lipid Metabolism
Nitrogen Metabolism
Nucleic Acid Metabolism
Membrane Transport
One-Carbon Metabolism
Oxygen Metabolism
Phosphate Metabolism
Protein Metabolism
Purine Metabolism
Pyrimidine Metabolism
Signal Transduction
Sulfur Metabolism
Users acquiring these distinct subsets will wish to impose a more refined
structure and some appropriate browsing mechanism. We realize that there are a
number of different ways in which the collection could be organized to support
ease of access. Our intent is to provide access in two of these ways, via the
EMP database and through the PUMA system on the WWW, and to provide mechanisms
for other groups to restructure and display the collection for their own
purposes.
Table 1
.
Number of charts for organism and taxonomic groups
|
Number of Charts
|
Organism/Taxonomic Group
|
|
667
|
Escherichia coli
|
|
531
|
Haemophilus influenzae
|
|
418
|
Homo sapiens
|
|
379
|
Mammalia
|
|
135
|
Rattus norvegicus
|
|
397
|
Rodentia
|
|
115
|
Saccharomyces cerevisiae
|
|
316
|
Aves
|
|
101
|
Ascomycotina
|
|
57
|
Salmonella typhimurium
|
|
56
|
Oryctolagus cuniculus
|
|
52
|
Bos taurus
|
|
51
|
Embryophyta
|
|
47
|
Pseudomonas
|
|
40
|
Bacillaceae
|
|
407
|
Mycoplasma capricolum
|
|
366
|
Archaea
|
|
29
|
Sus scrofa
|
|
26
|
Mus musculus
|
|
19
|
Pseudomonas putida
|
|
19
|
Clostridium
|
|
16
|
Pseudomonas
sp.
|
|
14
|
Tetrahymena pyriformis
|
|
14
|
Gallus gallus
|
|
13
|
Clostridium acetobutylicum
|
|
12
|
Neurospora crassa
|
|
12
|
Fungi imperfecti
|
|
12
|
Canis familiaris
|
|
359
|
Sulfolobus solfataricus
|
|
10
|
Pseudomonas fluorescens
|
|
10
|
Protozoa
|
|
350
|
Bacillus subtilis
|
Table 2
.
Number of enzymes for distinct organisms
|
Organism
|
Distinct enzymes in charts
|
|
Escherichia coli
|
434 enzymes
|
|
Haemophilus influenzae
|
335 enzymes
|
|
Homo sapiens
|
204 enzymes
|
A nomenclature for pathways
Early in the EMP project, it became clear that it was necessary to develop a
nomenclature of metabolic pathways, and rules for generating systematic names
were formulated. This was necessary to manage such a large collection (which
will grow to contain thousands of pathways and variations of pathways). This
nomenclature allows those wishing to restructure the collection or to write a
browser for the collection to have access to a compact representation of the
function of the pathway.
The systematic name contains the initial substrates, final products, the
function of the pathway, coenzymes, and cellular location of the pathway
enzymes. Every metabolic pathway record includes this characteristic systematic
pathway name. In addition, each record includes a shorter, but still
unequivocal, recommended pathway name. Finally, a set of common names for the
pathway are also encoded.
A brief description of the systematic name would appear as
Substrates-Products_Function_
(Coenzymes)_(Locations)_[Comment]
For example, one of the versions of the Entner-Doudoroff Pathway encoded in the
database is characterized by the name
D
-
glucose-pyruvate_catabolism_
(ATP,_NADP('+),_NAD('+),_ADP)_(cytosol)
while the recommended name would be
Glucose-pyruvate_catabolism_
[via D-glucono-1 ,5-lactone_6-phosphate]
and the common name would be
Entner-Doudoroff pathway
.
It should be noted that a number of pathways in the collection (those encoded
early in the project) do not strictly conform to the above conventions, and we
are attempting to correct omissions as quickly as possible.
PUMA, one means of access to the collection
We believe that this collection of metabolic pathways offers a powerful way to
organize access to other important categories of biological data. As an
illustration, we have added the pathways to the PUMA system developed at
Argonne National Laboratory to offer integrated access to biological data to
support interpretation of genomes (the URL for PUMA is given above). Within the
context of PUMA, the user has access to a general functional overview of
organisms in which the pathways have been broken into small functional
categories. By simply clicking on a functional category, one has access to the
set of pathways corresponding to the function. The enzymes in the pathways have
been connected to alignments, protein sequences, and related enzymatic and
sequence databases. The compounds connect to documents that summarize the sets
of pathways that utilize the given compound as a substrate, product, or
intermediate.
In addition to a general overview that attempts to organize the entire
collection of metabolic pathways, PUMA also offers access to over 200
functional overviews of specific organisms. Each of these overviews has been
constructed to include the metabolic pathways that have been identified for the
specific organism. We hasten to add that these overviews are far from complete;
indeed, they do not contain many pathways present in the literature. However,
we are making a concerted attempt to offer as complete a metabolic picture as
possible for those organisms being sequenced within the growing number of
genome initiatives.
As an example of what can be offered via the metabolic pathways, we encourage
the reader to access PUMA and examine the collection for
Escherichia coli.
When a specific functional category is accessed, the user is shown which
pathways implement the function. For each pathway, the enzymes for which
sequence exists are shown (links to acquire the sequence are provided),
alignments containing known sequences are indicated, and a list of the
unsequenced enzymes is provided. In the case in which a sequence does not exist
and for which the corresponding sequence does not exist for any other organism
(i.e., the sequence for the enzyme does not exist for any organism), the fact
is noted. The significance of including the EC numbers for each pathway becomes
obvious: they allow the pathways to be easily connected to the rich and growing
collection of databases containing enzymatic data (
3
-
5
).
Reconstruction of the metabolism of organisms from sequence data
One of the most significant applications of this collection of pathways will be
to support the analysis of the metabolism for organisms for which a substantial
amount of sequence data already exists. Indeed, now that several complete
genomes have been sequenced, the utility of metabolic overviews becomes
increasingly apparent. It is also worth noting as an aside that assignment of
function to coding sequences in a genome is a difficult task, and that access
to an accurate metabolic characterization of the organism or a close relative
often sheds light on the function of specific genes.
We are now completing collections of pathways to represent what is known of the
metabolism of
Haemophilus influenzae, Mycoplasma capricolum, Mycoplasma genitalium
and
Sulfolobus solfataricus.
As more genome sequences are completed, we will add their `constructions of
metabolism' to the collection. The process of reconstructing the metabolism of
an organism from sequence data involves using the sequence data to establish
enzymes that are known to be present, supplementing this list by using the
biochemical literature (since many sequenced genes have an unknown function,
while the literature often explicitly identifies the presence or absence of
specific functions), using what is known about phylogenetically close
organisms, and finally integrating these different sources of information. The
process of weighing different forms of evidence often reveals inconsistencies
and offers one of the important means for gradually addressing errors across
the public databases.
The reconstruction of the metabolism of these first complete genomes is a
challenging task that requires familiarity with a great deal of the biochemical
literature. However, since these early organisms were selected to ensure
phylogenetic and biochemical diversity, we believe that each step toward
formulating a more complete picture of their metabolism should dramatically
simplify the task for other related organisms. Indeed, we are working on tools
to support and automate sections of the task (
6
).
Comments on the format of the pathways
This collection of metabolic pathways originated as the `working notes' of
Evgeni Selkov. As an experiment, we spent a fair amount of effort converting
them to a set of relations that could be used in standard relational databases.
This approach appeals to those with a background in databases, and it certainly
simplifies a number of important issues relating to maintenance and
organization of the collection. However, the real choice is between keeping the
primary form of the data as charts (supplying software to produce the
relational representation) or keeping the collection as a relational database
(supplying software to produce the charts). In some sense the choices appear
equivalent. Our experience suggests that operationally this might not be
completely accurate. Certainly, the main curator of the collection finds it far
more convenient to encode and work with the actual drawings. It also seems
likely to us that, as other experts create and exchange encoded pathways, the
basic exchangeable unit should probably be drawings and not encoded tables.
We clearly need to move toward a situation in which the drawings conform rigidly
to a set of minimal standards that allow ease of conversion directly to a
relational format, along with tools that will render drawings in a variety of
forms for different purposes. To do this properly is more challenging than it
appears at first glance. We will attempt to facilitate moves in this direction,
but it must be emphasized that the charts often contain information well beyond
that given by the reaction equations, and this information must not be lost
(indeed, it seems likely to us that the relational encoding will naturally
evolve to contain a subset of the information contained in the drawings).
Availability
Downloadable demo versions of EMP are available via anonymous FTP. The
collection of metabolic pathways described in this article is publicly
available (free of charge) via anonymous FTP. For further information, send an
empty electronic mail message (or one containing the single word INFO) to
emp_info{at}biobase.com.
Those who do not have access to electronic mail may write to Biological
Databases Inc., 2004 South Wright Street, Urbana, IL 61801, USA.
ACKNOWLEDGEMENT
This work has been supported in part by the US Department of Energy.
REFERENCES
1 Selkov, E., Goryanin, I., Kaimachnikov, N., Shevelev, E. and Yunus, I. (1990) In Glaeser, P. (ed.), Scientific and Technical Data in a New Era. Hemisphere Publishing Corp., pp. 22-27.
2 EMP User Manual: A Guide to the Enzymes and Metabolic Pathways Database (1995) Release 1.0, September.
3 Bairoch, A. (1994) Nucleic Acids Res., 22, 3626-3627.
4 Bairoch, A. and Boeckmann B. (1994) Nucleic Acids Res. 22, 3578-3580. MEDLINE Abstract
5 Suyama, M., Ogiwara, A., Nishioka, T. and Oda, J. (1993) Comput. Appl. Biosci., 9, 9-15. MEDLINE Abstract
6 Gaasterland, T., Maltsev, N. and Olrerbeek, R. (1995) In Proceedings of the Second International Meeting on Integration of Molecular Biological Databases, Cambridge, England, July.
Return
*
To whom correspondence should be addressed
