Differences in genetic database information could lead to misdiagnosis

Inconsistencies among a growing number of genetic databases could lead to misdiagnoses of genetic disorders by health care providers, according to a study that appears in Cytogentic and Genome Research.

The study, conducted by RTI International researcher Dr A Jamie Cuticchia, compares the cytolocations (or locations on the human chromosomes for specific traits) for genes listed on two popular catalogues Ensembl and the online version of Mendelian Inheritance in Man.

The study found that at least 18% of the records analysed have at least one cytogenetic band discrepancy between the datasets. If the level of inconsistencies in these two data sets reflect the level of inconsistencies among other biological data, it creates a situation where if more than seven data sets are merged, there is less than a 50% chance that any record obtained will be the record the researcher expects.

"Geneticists rely on cytolocations as the primary reference point for locating human genes," said Cuticchia, principal investigator for the study. "We need genome sequencing and genetic mapping to match to ensure that doctors and geneticists are referring to the same genes at the same locations."

The growth of genomic information in recent years has led to an increase in the number of available databases as well as increasing number of inconsistencies among data, decreasing the reliability of information when datasets are merged.

"The discrepancies between the various datasets will significantly impact the usability of the data and could delay in the development of new treatments," Cuticchia said. "Accurate genetic cytolocations are of importance to geneticists who discover or confirm clinical applications for genetic disease gene candidates. Discrepancies in cytolocations among databases delay research and ultimately, treatments."

According to RTI researchers, one of the problems in keeping biological data consistent among databases is the inability of groups to easily and directly input their own modifications to the data into the databases.