July 10, 2015
A study published in Genetics in Medicine describes the problems associated with using the current variant databases to reach a diagnosis. It is well known that Marfan syndrome (MFS) – an autosomal dominant disorder caused by variants in FBN1 which expresses a protein important to connective tissue – is diagnosed using FBN1 mutation test according to its diagnostic criteria. The authors report that there are 23 common variants of FBN1 registered in ESP6500 which are classified as causing MFS in the Human Gene Mutation Database (HGMD). The researchers evaluated both database information and papers cited in the databases against defined clinical diagnostic criteria for MFS known as the Ghent nosology.
They found that none of the four well known variant databases they consulted (HGMD, UMD-FBN1, ClinVar, and UniProt) showed that the 23 variants of FBN1 had clear associations with the disease. According to the authors, many papers referenced in these databases contained phenotypic data associated with specific variants but lacked evidence that patients had received an accurate MFS diagnosis. The authors warn that researchers and doctors should exercise extensive caution while using variant databases as many of them are uncurated, inaccurate and not entirely reliable.