Reference Genome Refresh

The use of whole-genome sequencing to diagnose rare disease is on the rise. In such cases, an individual’s genetic map is compared with the reference genome – a “standard” version, if you will. Although many of the differences between the two are not significant, some divergent sequences can be the drivers of rare disease. However, new evidence uncovered by a team from the Karolinska Institutet in Sweden indicates that the human reference genome might be outdated – at least in regard to its diversity (1). We spoke to Jesper Eisfeldt, a doctoral student in the Rare Diseases research group and first author of the paper, to find out more.

Tell us about your discovery…

We reanalyzed 1,000 previously sequenced Swedish genomes (2) to isolate the regions that are not present in the reference genome. We discovered over 61,000 novel sequences – a volume equivalent to approximately one whole chromosome – affecting over 80 genes, 12 of which are linked to various diseases. After comparing these novel sequences to the genome library for chimpanzees, Icelanders, and the African population, we discovered that they are highly conserved and widely distributed in the human population. This showed us that the human genome is more heterogeneous than previously thought and, as a result, we need to update our reference genome.

What is the potential impact of this lack of diversity in the reference genome?

Sequences missing from the reference genome could be of clinical relevance. For example, they could contain regulatory elements. In some cases, we are unable to fully understand large variants partly located in these missing or wrongly positioned sequences. When comparing our reads to the reference genome, we are at risk of positioning the missing sequences in the wrong place, producing noise and detecting false-positive variation. A more complete reference genome will give us cleaner and more reliable results.

How can the reference genome be changed?

The current reference genome, GRCh38, was made available in 2013. Although continuously improved (the latest patch was released in March 2019), it is based on previous versions of the human genome and largely produced though hierarchical shotgun sequencing. We believe that a graph-based reference genome is the way forward, because graphs can be used to represent multiple variants of the same sequence. In the current reference genome, which is linear and haploid, it is tricky to represent complex and diverse variation.