NucleaRDB: Automated extraction of point mutations from the literature

This section provides point mutations extracted from the literature in an automatic manner

Point mutations (PMs) were extracted using MuteXt (Horn et al. 2004). MuteXt extracts PMs from full text papers (or Medline abstracts when full texts are not accessible).
MuteXt does not distinguish 1)single PMs (e.g. F234H) from multiple PMs (e.g. F234H/F255C/G277V); 2) PMs that are described by the authors from PMs that are only cited and published elsewhere. MuteXt misses PMs that are described in figures and PMs with a too funny nomenclature (e.g. FH234, H234 instead of F234H).
Residue numbering: All the point mutations are numbered using the Swiss-Prot residue numbering. When appropriate, the authors' numbering is indicated in the pages that describe individual point mutations.
General residue numbering: We also use a general numbering in order to compare residue positions between receptors. More details on the NucleaRDB numbering is available here.

The complete mutation database contains 4490 PMs extracted from 1211 articles (2921 articles have been processed in total). 2617 PMs were automatically validated (positives) and 1873 PMs are considered negatives. Only the validated PMs are published on this web site at the moment. I remove false positives when they are identified. 160 false negatives (so far) are not yet available via the NucleaRDB.

Access the data via

• The list of extracted point mutations
• The list of receptors
• A simple query system

• or more mutation data via the NRMD query system