It has been proposed that DNA damage from endogenous sources gives rise to 20000 lesions per mammalian cell per day, most of the lesions being deaminations, spontaneous hydrolysis of the N-glycosidic bond, alkylations, and damage by reactive oxygen or nitrogen species and lipid peroxidation products (9–13). Lesions are also caused by errors in DNA metabolic processes, including the formation of single- and double-strand breaks, the collapse of replication forks, and the introduction of modified nucleic acid bases during DNA replication. Counting all together, daily 10¹⁶–10¹⁸ repair events occur in a healthy adult man (10¹² cells) (14). Despite the protection provided by these mechanisms some of the damage escapes repair, and in consequence leads to mutations, ageing and various diseases, including carcinogenesis and neurodegeneration (15–19).

DNA repair is a very complicated process, involving many factors. For instance to date, 168 genes that encode proteins involved in DNA repair have been identified in the human genome (18–20). They are involved in diverse processes, starting from detection of a damage site in the DNA, through several steps of enzymatic transformation of the damaged DNA, to recombination and signaling to stop the cell cycle or initiate apoptosis. Another form of dealing with the DNA damage is lesion bypass, which facilitates continuation of replication even when irremovable modifications occur, but does not guarantee proper recreation of the original sequence and frequently leads to mutation generated by translesion synthesis (TLS) polymerases.

The numerous chemical and structural transformations leading from damaged DNA to repaired DNA can be described as pathways, comprising series of reaction steps. Currently known DNA damage signaling (DDS), repair and damage tolerance pathways can be divided into eight categories: Each of these pathways can be represented as a series of enzymatic transformations between different DNA structures, catalyzed by a dedicated system of proteins. It must be emphasized that DNA repair pathways are connected to each other, i.e. they can share some steps and/or proteins involved (14). As a consequence, DNA repair proteins rarely work in isolation in the cell, and their activity is dependent on other components of DNA repair systems. Therefore, knowledge of both the entire DNA repair systems and their components is critical to our understanding of how cells controls and repairs the constantly occurring damage of their genomes. Many of the proteins involved in DNA repair are very well-described (including substrate specificity, kinetics, mechanisms of action and 3D structure in complex with the substrate and/or the product of its activity). Information about different factors involved in DNA repair is, however, scattered in the literature and thus far there has been no resource to organize information on DNA repair at the systems level. Moreover, there are still processes, for which an enzymatic activity is known or suspected to exist, but the genes/proteins/enzymes proteins have not been characterized yet. As an example an enzyme that is capable of removing 5-hydroxymethyluridine, a product of thymine oxidation in human cells must exist, but to date has not been identified yet (21).

DAMAGE: We collected information about 85 different types of damage in the DNA (as of October 14, 2010). Many of them describe general classes of damage events such as single-strand breaks or base loss that are independent of the local sequence). About 60 chemical compounds that cause DNA damage were connected to the according types of damage. Of all lesions, 36 could be connected to a single molecular structure (e.g. point mutations and nucleotide modifications such as 1-hydroxypropyl-adenine). Each type of damage is described on its own sub-page, which includes information about the potential source (e.g. spontaneous formation, intermediate in some DNA repair process, etc.), proteins that may recognize its presence in the DNA, keywords that facilitate analyzing its context, and literature links. For 36 types of damage with unique chemical structures, REPAIRtoire displays the structure in 1D (using a SMILE code), 2D and 3D (with the JMol JAVA applet), and provides atomic coordinates for download in the .mol format.

PATHWAYS: This menu provides access to data about eight pathways (DDS, DDR, BER, NER, MMR, HHR, NHEJ, TLS) from three model organisms: Escherichia coli, Saccharomyces cerevisiae and Homo sapiens. These pathways are represented as graphs visualized with PyGraphviz (http://networkx.lanl.gov/pygraphviz/), in which the nodes represent DNA states, and the edges represent the reactions between them, e.g. enzymatic reactions. All edges of the (sub)graph, i.e. arrows that connect the images, are hyperlinked to static ‘reaction’ windows comprising one or more panels that display basic information about the selected reaction. All nodes of the graph, e.g. DNA–protein complexes at various stages of the repair process, are hyperlinked to static windows comprising detailed information about the given stage of DNA repair. All protein components of each state/complex are also hyperlinked to individual protein pages.

PROTEINS: As of October 14, 2010, REPAIRtoire stores information about 69, 78 and 154 proteins from E.coli, S.cerevisiae and H.sapiens, respectively, and their genes that can be assessed either via pathways or directly from the `PROTEINS’ menu. In the process of manual data curation, we collected the available information concerning alternative gene and protein names and amino acid sequences available in the NCBI databases (22), 3D structures available in the protein data bank (23) and various features available in other databases (e.g. information about the enzymatic function, presence of isoforms, cellular/tissue and subcellular localization, together with links to the relevant database entries). Currently the DNA repair protein data set encompasses only E. coli, S. cerevisiae and H. sapiens, but will be expanded in the future and may eventually comprise all orthologs of the functionally characterized enzymes identifiable in fully sequenced genomes.

DISEASES: Thus far, we have compiled information about 40 diseases caused by the mutations in 32 genes linked to defects in DNA repair proteins. This data set is presented as a table with hyperlinks to the proteins concerned, and to the relevant entries in the Online Mendelian Inheritance in Man (OMIM) database (22). Each disease has its own subpage with a succinct description and additional links (e.g. keywords). Reciprocal links to diseases are also available in each protein field.

KEYWORDS: This menu provides quick access to the most common keywords used to annotate the database entities according to biological processes and activities such as: DNA repair pathways, the response to DNA damage, the cell cycle checkpoint control, the DNA N-glycosylase, the DNA N-glycosylase/AP-lyase, etc.

We would like to emphasize here that there is no need to register or to log in to view the content of the database. However, the profile and login features have been provided for collaborators and `super users’ who are interested not only in viewing, but also editing the content of the database. Creating an account and loging into the database give the user access to the admin site of the database, but also uncovers the wiki-like pages of the entries. By entering the administration site it is possible to add new data, delete information, edit and correct mistakes. Editing information about proteins, genes, diseases and types of damage is also available via wiki-like pages, from the pages of particular database entries. Users can also add comments and new references to existing records.

The image drawing tool (accessible via a `draw a picture’ link in the main menu) system has been developed to illustrate all steps of DNA repair pathways as protein–DNA complexes, in which proteins are displayed in the textbook-like format of ‘potato models’ (ellipsoids). However, it can be also used as a standalone tool (without any connection to REPAIRtoire) to create images of protein–protein or protein–DNA complexes. The drawing engine uses the SVG format provided by the W3C consortium, and enables exporting the image in the JPEG format. The SVG format enables resizing the images without the loss of the quality and makes it possible to modify them with external tools for vector graphics processing, e.g.Inkscapeor others free or commercially available software.