In the news
Self-reported medical data supports genetic disease links
|Study:||Efficient replication of over 180 genetic associations with self-reported medical data.|
|By:||Tung J.Y. et al. (11 authors total)|
The study authors undertook an evaluation of an online recontactable research cohort with self-reported medical information for use in genetic association studies.
The researchers from genetic testing company 23andMe used their customer database to ‘recruit’ over 20,000 participants of European ancestry with data on 600,000 single nucleotide polymorphisms (SNPs). Participants answered online research surveys in order to provide phenotype data on 50 medical conditions. The researchers tested an extensive list of SNPs identified from published genome-wide association studies in order to replicate previously identified associations across many diseases.
The 23andMe researchers were able to replicate 75% of the associations they expected to (based on power calculations) including conditions such as type 2 diabetes and prostate cancer. Many of the associations not replicated in this study did show the same direction of association although several of them have already been replicated elsewhere in the published literature.
The study authors state “that web-based collection of self-reported data on medical phenotypes is an efficient and effective method for phenotyping a large cohort of individuals, as evidenced by our ability to replicated a high percentage of associations across a wide range of conditions.”
Having recently published a proof-of-principle study (see previous news) the study authors further tested the idea that internet-based phenotyping can be used to easily replicate identified associations in a fast, efficient and effective manner. This method seemed to work well for many conditions in individuals of European ancestry (in whom the majority of published studies are conducted). A trade-off of this increased speed in replication is that the simplicity of the research survey used may not be sensitive enough for appropriately phenotyping psychiatric conditions, something the authors acknowledge in their discussion. One key strength of this methodology could be the relative ease with which individuals can be recontacted, possibly to conduct more in depth phenotyping or to increase the power of future studies.