Did you learn something by using the Pain Genes Database?
Please remember to cite our companion manuscript:
LaCroix-Fralish, M.L., Ledoux, J.B. and Mogil, J.S. The Pain Genes Database: an interactive web browser of pain-related transgenic knockout studies. Pain, 131:3.e1-3.e4, 2007.

Here's how it works: 

To identify published manuscripts describing results of pain-relevant knockout studies, we have and will continue to perform searches of the MEDLINE/PubMed database.  Two specific searches are conducted.  The first uses the U.S. National Library of Medicine’s controlled indexing vocabulary, MeSH, with the terms “Pain Measurement” AND “Mice, Knockout.”  The second uses the search string:  “(mutant OR knockout OR knock-out OR deficient OR transgenic) AND (mice OR mouse) AND (pain OR *nocicepti* OR hyperalgesi* OR allodyni*) NOT review.”  We then peruse each individual abstract (and where in doubt, the full text of the manuscript) to identify papers describing experiments with the following characteristics:

  1.  Testing of a spontaneous or genetically engineered mutant mouse with null expression of a single gene.  Not included are transgenic mice overexpressing a gene,  as this would complicate understanding of the “sign” of the phenotype, nor double (or triple) knockouts, as this would complicate genotype-to-phenotype attribution.  In cases where a spontaneous mutation and a transgenic knockout of the same gene exist (e.g., the leaner mutant and the Cacna1a knockout), only the transgenic knockout data are considered since these data are generally more recent.  Cre/lox-based conditional knockout data are included only in cases where the “first-generation” null mutation proved lethal; the database indicates the restricted spatial expression of these transgenes.  Similarly, data from heterozygotes are included only in cases where the homozygous null mutant is lethal.

  2.  Testing of the mutant mouse on a behavioral assay of acute or tonic nociception, injury- or stimulus‑induced hypersensitivity (i.e., allodynia or hyperalgesia), or drug- or stress‑induced inhibition of nociception (i.e., analgesia).  A decision was made not to include papers solely featuring pain proxies, including electrophysiological, gene expression, immunohistochemical and/or neurochemical data.  We also decided to exclude papers related to anesthesia and opioid tolerance and dependence/withdrawal, although acute nociceptive and analgesic data from those papers are considered if supplied.

  3.  The claim of at least one statistically significant difference, in either direction, between the mutant mice and their simultaneously tested wildtype (+/+) controls on any of the phenotypes described above.  If a mutant was tested but no significant differences found in any phenotype described above, the paper is not included.  If at least one significant difference was found, the positive and negative findings from that paper are included in the database.  The exception to the main rule is the inclusion of studies featuring only null results for knockout mice having previously been shown to display a positive result on nominally the same phenotype. 

The Pain Genes Database has two levels of exploration:  across-gene and within‑gene.  The across-gene level, the PainGenesdbSelector, is encountered first.  All genes in the database can be accessed and sorted by their gene name, protein name, common names and acronyms, or genomic position (by navigating a graphic representation of the mouse genome).  The gene and protein names can be selected from an alphabetical list, or by typing a text string into a search box.  Common names and protein acronyms can be found by typing the text string; note that Greek letters have been replaced by the analogous Latin character.  We expect that a more common approach will be to search for groups of genes based on their protein functions.  To this end, we have assigned entries in three categories—Cellular Process, Function, and Subfunction—to each gene/protein.  This assignment was loosely based on Gene Ontology (GO) categories.  There are only five Cellular Process levels:  Cell Division, Cell Signaling, Cell Structure, Gene Regulation and Metabolism.  By contrast, there are multiple (and potentially unlimited) levels of Function and Subfunction, with the latter category serving either to increase specificity or to accommodate bridging between different biological functions.  We invite feedback from the research community to increase the accuracy and usefulness of these categories and the mapping of proteins onto them.  Finally, one can search for groups of genes by their “summary” phenotype (i.e., :  mutant more sensitive, :  mutant less sensitive, :  no difference, :  not tested, ?:  contradictory data), an arbitrary decision made  after looking at the full published data in three main phenotypic categories:  Nociception, Hypersensitivity, and Analgesia, although in many cases the actual phenotype is considerably more complex.  If more information about the gene itself is desired, users can  link from the Pain Genes Database directly to the Entrez Gene database  entry.

Once a gene or set of genes has been chosen, the user can jump to the within-gene exploration level, the PainGenesdbBrowser, in which the phenotypic data from all published papers relevant to each gene can be accessed.  On this screen, one can quickly browse the data from individual papers in which each mutant has been tested, or link directly to the MEDLINE/PubMed abstract.  Each of the three main phenotypic categories is now broken down into sub-phenotypes based on the specific tests performed.

In every case, one of the following entries will be seen:  (i.e., mutant is significantly more sensitive than wildtype), (i.e., mutant is significantly less sensitive than wildtype), (i.e., no significant difference between mutant and wildtype sensitivity), or a blank space (i.e., not tested in that study).  “Mousing over” the // entry will reveal a text box with relevant parameters of the behavioral assays (e.g., hot‑plate temperatures, von Frey filament presentation order, formalin concentration), drug names and dosing, and injury models.  Note that where latencies are provided in seconds, these refer to baseline latencies of wildtype mice.