Metabolic Pathway Databases: A Word of Caution

Abstract

Metabolic pathway databases collect and organize knowledge on metabolism that has been gathered over the course of decades of research. It is, however, far from trivial to accurately represent all this knowledge in a format suitable for a wide range of computational analyses. Nevertheless, many research groups have taken up this challenge, inspired by the great value and potential of pathway databases. Nowadays, these databases are routinely used in the interpretation of omics data. They have also proven to be powerful resources for various other types of analyses, including the identifi cation of potential drug targets and in silico phenotype prediction. The number of metabolic pathway databases continues to grow, describing the metabolic network for an increasing number of organisms. For Homo sapiens and several other organisms, such as S. cerevisiae and A. thaliana , there are even multiple databases that describe their metabolic network.A word of caution is, however, warranted when working with metabolic pathway databases, because no two are alike. Databases have different ways of representing a metabolic network in a digital format and they defi ne concepts like a pathway differently. Moreover, the breadth and depth with which a network is covered varies. Several comparisons have also shown that the consensus between databases that describe the metabolic network of the same organism is limited. Furthermore, each database offers users a different way to extract the knowledge represented. At the root of these differences lies the main intended use of the database, the strategy used to reconstruct the metabolic network and the degree of curation.Taken together, these differences should be borne in mind when working with metabolic pathway databases, especially when one needs to choose between several databases that describe the metabolic network of the same organism. The choice of a particular database could infl uence (the interpretation of) the results of the analysis at hand. In this chapter, we provide further insights into metabolic pathway databases to understand their individual strengths and limitations, which is important to be aware of for any essential input of an analysis. Ultimately, these databases should be integrated and consolidated to arrive at a more complete and biologically accurate description of the metabolic network. This will, however, require a broad community effort.

Metabolic pathway databases and model repositories

“…These reconstructions aim to integrate metabolic entities (genes, enzymes, reactions, and compounds) for a better understanding of the cellular structure and functions [9]. Well curated and experimentally tested reconstructions have been applied in wide variety of areas such as identifying essential gene/ reactions and drug targets [5], studying protein-protein interactions, producing and increasing the yields of important industrial chemicals [5,10,11]. Therefore, it is of great value to assess and provide an insight view of the existing databases and pathway resources so that to help the researchers in the field choose the most suitable one according to the information and data pertinent to a particular research.…”

“…A comprehensive list of the pathway resources from different categories of life sciences is available at http://www.expasy.org/links.html and http://www.pathguide.org. Information from pathway resources has been extensively applied in in-silico reconstructions and analysis of biological models representing vital cellular systems such as signaling, regulatory and metabolic networks [5]. Accordingly, hundreds of metabolic pathway models have been reconstructed and represented in the form of predictive mathematical models for those organisms with their genome sequenced [6][7][8].…”