The FH mutation database  is based on the recently described Leiden Open (source) Variation Database (LOVD) system  which, under the auspices of the Human Genome Variation Society, is rapidly becoming the standard for gene-specific databases. The LOVD system was designed to be flexible so that it could be easily adapted to the needs of a wide range of locus specific databases. Additional flexibility comes from the platform-independent design and the use of PHP and MySQL open source software. The design is gene-centered and modular, and focuses on the collection and display of DNA sequence variations. The LOVD platform is also easily extendable to include summary clinical data. In addition, all database contents may be downloaded in flat text format and imported into a spreadsheet program for further analysis. The LOVD database software is now in a new version, LOVD2, which has many additional features.
Researchers interested in contributing sequence variants to the database will be asked to register, (submitters can access and edit their personal data at any time), after which they can contribute the variant online. During submission researchers are asked to provide those data that are deemed absolutely essential for mutation databases by the Human Genome Organization Mutation Database Initiative. These include a patient ID, an exact molecular description of the variant (DNA-level), and details about the source of the material and detection method used. Newly submitted variants are given a unique identifier, as recommended by Claustres et al . After the curator's approval the new variant is added to the database and all connected web pages are updated instantly. Currently (Jan, 2008) all database content is derived from published, peer-reviewed literature. We anticipate that directly submitted content will become an increasing proportion in the future. As the submitted data are in the form of a submitter deduced mutation description, the curators can only check the feasibility and consistency of the description and not any original data.
DNA variation in populations is generally benign and as such should be termed a 'variant' or 'SNP'. The term 'mutation' is reserved for cases were there is a reasonable suspicion that the variant is deleterious. DNA variants submitted to the FH database should be described in accordance with the recommendations of the Human Genome Variation Society (HGVS), update August 2004, and it is strongly recommended that authors consult the most up-to-date nomenclature guidelines, which can be found online at the HGVS website. Unfortunately the mutation nomenclature currently used in the field of FH research is rarely in agreement with HGVS guidelines. Reports contain a variety of annotations, and many are based on the annotation of the cytosolic protein, the function of which is currently poorly defined . One of the aims of the HGVS is to establish a consistent nomenclature across all human genes. The description of FH protein variants based on the cytosolic protein is often accompanied by the use of a truncated transcript to describe cDNA variants. Other authors use the cytosolic protein numbering system, while applying the full-length coding transcript to describe cDNA variants. This results in the loss of the simple approximation of 3 to 1 between the numbering. The HGVS-recommended cDNA numbering starts from the first ATG of the full coding sequence. Protein reference sequences should represent the primary translation product, not a processed mature protein, and thus include any signal peptide sequences Currently many variants are present in the literature under several titles, and the correct identification of intron variants is especially problematic. No reliable overview of known FH variation can be achieved without a consistent nomenclature, which includes the appropriate numbering systems at both the DNA and protein level. We now recommend that all authors implement the recommendations of the HGVS when describing future FH variants.
The FH mutation database will accept all variants and if necessary assist in assigning the correct nomenclature. In light of the inconsistency of the recent FH mutation nomenclature, variants have been reassigned, where necessary, and an extra column has been generated clearly stating the originally reported description(s) of the mutation at the protein level, and where equivocal also at the cDNA level. Inclusion of sequence variants in the FH mutation database does not imply that there is convincing evidence for pathogenicity. Please refer to the disclaimer on the website.
Within the FH database, all variants that disrupt the reading frame, affect highly conserved residues (conserved by sequence alignment in at least Saccharomyces cerevisiae) or disrupt the consensus donor or acceptor splice sites (GT-AG), and are not found in healthy controls, should be suspected as potentially pathogenic. Non-conserved missense variants and potential splice site mutations that do not disrupt the consensus splice sites, are considered to be SNPs or rare variants of uncertain significance. Ideally, several lines of evidence should support publications which present mutations as "pathogenic", including screening DNA from a panel of 50–200 healthy individuals, describing the nature of the amino acid substitution (conservative/non-conservative) and the significance of the position in the protein (evolutionarily conservation or known functional domain). It is important to note whether the mutation has previously been reported, if it has been found in several families, and if it segregates with the disease within the family.
Many FH variants have been analysed for effects on enzymatic function, and this data has, in many cases, allowed the classification of missense variants as pathogenic. In addition a SIFT (Sorting Intolerant From Tolerant) analysis  of missense FH variants has been included. A high score (above 0.05) indicates that an amino acid is poorly conserved (alignment of 337 Uniprot proteins), indicating caution in assignment of function to the variant.
However none of these factors can be seen as definitive and each variant must be considered on its merits. Unfortunately, most mutations are currently reported without this accompanying analysis, and many have been identified in a single case or family. Thus caution should be exercised when attempting to derive clinically relevant information from the database, and users must carefully weigh all the evidence in the database and any additional data.
The FH gene is very highly conserved, showing 67% protein identity between humans and S. cerevisiae with the few coding SNPs described in dbSNP present in the HapMap population as very rare alleles (the minor allele being undetected). Thus, current knowledge suggests that most missense changes will be pathogenic. Polymorphisms, including intron variants, synonymous (silent) variants, nonsynonymous missense variants found in a healthy control panel and potential non-consensus splice site mutations but without evidence for transcript rearrangements are included in the database as such unless accompanied by clear evidence of pathogenic potential.
An abbreviated description of the FH related syndrome is given under "disease". Refer to the database notes for an explanation of these abbreviations. Further information can be found under "remarks", including the country of origin of the patient or study, together with other details such as the number of healthy controls tested for the variant, and any other supporting evidence.