Research Highlights

Strict statistics: Overlooking disease associations

Published online 19 July 2017

A commonly used statistical approach is missing associations between DNA variations and disease.

Nadia El-Awady

A variety of statistical approaches are used to investigate associations between small changes in the DNA-makeup and specific diseases. The researchers have validated a statistical approach that produces a higher number of results1.

The international team of researchers, including one affiliated with the Lebanese American University in Beirut and another with King Abdulaziz University in Jeddah, analysed data from the UK Biobank from more than 10,000 people with coronary artery disease. 

The data include the genetic constitution for each person, or their genotypes, in addition to self-reported and physician-documented information. They used different statistical approaches to analyse these data, comparing the results with samples from other large studies.

“We hypothesized that the approach that has been adopted as standard is unnecessarily cautious and so delivers few confirmed results. So, instead, we used and validated a different statistical approach that accepts that only the large majority of individual results needs to be correct,” says geneticist Hugh Watkins of the University of Oxford, one of the study’s lead authors.

By using this statistical method, called ‘false discovery rate’, the team identified more than 300 DNA variants they expect are linked to coronary artery disease. Further investigations of these variants could lead to a more comprehensive understanding of disease development and thus to new preventive and therapeutic measures. 


  1. Nelson, C. P. et al. Association analyses based on false discovery rate implicate new loci for coronary artery disease. Nat. Genet. (2017).