Researchers from across the globe have created the world’s largest publicly available dataset of genetic variants – changes in the DNA that can sometimes lead to disease.

This resource called the Genome Aggregation Database (gnomAD) includes information from over 140,000 people from populations around the world, and expands on the work of the 1000 Genomes Project; the first large-scale international effort to catalogue human genetic variation.

Today sees the publication, in Nature journals, of the first major studies using this database. This landmark collection from gnomAD consortium scientists includes two papers led by researchers  at the MRC LMS and Imperial College London. These studies reveal previously unknown genetic variants that lead to disease, as well as insights into a potential drug target for the degenerative condition Parkinson’s disease.

Dr James Ware, Head of the Cardiovascular Genomics & Precision Medicine group at the MRC LMS, co-author on three of the papers and a member of the gnomAD consortium said: “Reading genomes has become routine in both scientific research and healthcare. However, there are still enormous challenges in understanding what is written in our DNA, and in using that information to improve the health of patients.

“The gnomAD resource is proving hugely valuable to understand which genes are important in human health and disease, and to understand which specific variants in those genes cause problems, and it is already in daily use in genetic testing laboratories around the world.  This collection of papers highlights just some of the many uses for this powerful resource, which will become even more informative as the dataset gets bigger.”

“The gnomAD resource is an enormous team effort. More than 100 different research teams have shared data from over 140,000 people.  This sort of collaboration and sharing allows work that no single team could manage in isolation”.

Tools for better diagnosis of rare diseases

By looking across the large number of individuals in the gnomAD database, researchers can observe different patterns of variation in different genes. Some genes are quite commonly disrupted, while others are very rarely altered in healthy people, suggesting that variation is likely to be harmful. These huge datasets can contribute to helping clinical geneticists to more accurately determine if a given genetic variant might be protective, neutral or harmful to a patient.

One of the MRC LMS and Imperial-led studies in this collection explores the impact of genetic variants arising in the so-called untranslated regions of the genes. These regions lie just ahead of where a gene begins at the point at which the the molecular machinery in the cell starts to read the DNA code of that gene, and convert it into protein. These untranslated regions do not get converted into proteins themselves, so any defects in these areas can be much more difficult to spot, and are not the traditional focus of clinical diagnosis.

Dr Nicky Whiffin of MRC LMS and National Heart and Lung Institute (NHLI), Imperial College London, James Ware and colleagues explored the impact of genetic variants in these regions in 15,708 individuals. They found that changes in these regions could trick a cell to start reading the gene in the wrong place, meaning that a potentially vital protein would not be made by the cells, leading to rare diseases. One such condition that can be caused from these types of DNA changes is neurofibromatosis; a rare condition that causes tumours to form on nerve tissue.

Dr Whiffin shared some exciting developments as a result of this study:

“This new genetic library has helped us to pin-point specific variants that are under strong selection and can lead to certain diseases. Since the work was first release as a pre-print on bioRxiv, I have been approached by multiple people who have used it to identify new genetic diagnoses for previously undiagnosed patients. Finding the genetic cause of a disease can be hugely valuable to a patient – allowing family screening to find additional at-risk family members, driving treatment approaches and sometimes informing reproductive decisions. By finding these new diagnoses we can give answers, and sometimes hope, to these patients and their families. It is fantastic to see how this work is already having a translational impact. I am excited to continue this work looking at non-coding genetic variants that cause rare disease as I set up my own research group in Oxford later this year.”

Guiding drug development for conditions such as Parkinson’s

In a second paper, MRC LMS and Imperial colleagues, along with others from other institutes including the Broad Institute, 23andMe and the Michael J Fox Foundation, used the gnomAD database to suggest that a potential treatment for Parkinson’s should be safe in humans.

They explored the role of LRRK2, an enzyme called a kinase. Genetic variants that increase the activity of this enzyme are a known cause of Parkinson’s disease. As a result, many drug companies are interested in developing treatments that stop the action of LRRK2. Initial studies have, however, shown worrying side-effects in the lung, kidney and liver for these types of treatments.

In this study, researchers used three population-scale genetic databases; gnomAD, 23andMe and the UK Biobank, to look directly at data from over 4 million individuals to assess the consequences of reducing the expression of the enzyme LRRK2. The team found that these so-called loss-of-function genetic variants do not lead to severe organ dysfunction in humans. This is promising news for Parkinson’s disease patients and for drug companies who are currently progressing drugs against LRRK2 through clinical trials.

Dr Whiffin commented: “Large genetic databases with linked health information can give hugely powerful insights into the likely impact of certain drugs. By using nature’s own experiments, we can study people who naturally have reduced amounts of LRRK2 protein. If we see that genetic variants that naturally reduce the amount of protein in our body do not result in severe disease or reduced organ function, we can be more confidence that targeting that protein therapeutically will also be safe. This is only one piece of the puzzle, however – ultimately we need the results of clinical trials to really tell us about the safety and effectiveness of a drug in humans, but this is promising news for patients.”

 

‘Characterising the loss-of-function impact of 5’ untranslated region variants in 15,708 individuals’ was published in Nature Communications on 27 May. Read the publication here.

‘The effect of LRRK2 loss-of-function variants in humans’ was published in Nature Medicine on 27 May. Read the publication here.

 

 

Header image credit: Abby Chung, image sourced from Pexels.com