Knowledge of what’s “normal” is crucial for almost all medical tests. A raised serum cholesterol is meaningless without knowing how much it is above the normal range. An X-ray is uninterpretable without years of study to know what the image should look like. Even commonplace measurements like weight and height require reference ranges in healthy, age-matched individuals to be medically informative.
So what’s normal in a genetic test? A genetic test result isn’t a continuous variable, where a single measurement can be compared against a normal distribution, but a discrete change in the genetic code at specific locations.
Nonetheless, we need information about how likely those types of changes are to occur in the healthy population in order to interpret genetic changes in an individual patient. So to understand a single genome, many many thousands of other genomes are needed for comparison.
Everyone’s genome has around 4 million differences
Everyone has around 4 million locations in their genome where they may differ from another person. The genes comprise only ~2% of the genome but everyone still has around 22,000 places in their genes where they may differ from another person.
The majority of these differences are substitutions of single letters of DNA code. But there are also insertions and deletions of small pieces, or sometimes, large pieces of DNA. The vast majority of these genetic changes are very common, shared by many individuals across a particular ethnic group. Less than 5% per genome are rare or ultra-rare, shared only amongst immediate family members.
We can tolerate most gene changes
However, there is lots of rare variation across populations, and even the most constrained portion of the genome – the genes – contain an average of 1 variant every 8 base pairs (that we know of so far).A decent number of these gene changes will have an effect on gene function but do not have any medical impact. In part this is because we have two copies of every gene so affecting the function of one copy isn’t problematic; the working copy can make up for it. But some genes seem to be completely dispensable and we can tolerate both copies of the gene not working.
Overall, almost all of the 4 million changes in a person’s genome are benign and have no medical relevance!
Every genome contains 40-100 new changes
Most genetic changes are inherited from our parents, passed down through generations. But every individual’s genome also contains 40-100 new changes, not present in either parent, that occurred spontaneously as a by-product of reproduction. Of these, 1 or 2 will likely be located in a gene. In some cases, these new changes can be catastrophic and result in severe developmental disorders, but in most cases, they are benign.
Sorting out benign and medically relevant variants is hard
As we showed in a previous blog, sorting out benign variants from the ones that are medically relevant can be very difficult. Even in genes that have been shown to cause various serious medical conditions, it is a problem. This is because such disease genes still contain plenty of normal, benign changes in the healthy population, so picking out which changes cause disease, and which don’t, remains a challenge.
Unfortunately, it is a problem that is only getting bigger, as more and more tests, encompassing more and more genes, and detecting more and more variants are being done.
Population genetics is key to genetic medicine
As we move into an era where sequencing genes and genomes becomes routine in medicine, detailed knowledge of population variation both within specific genes and across whole genomes is crucial for interpreting individual patient genetic data.
Data sharing is key to this endeavour, and large collections of data such as ExAC have already successfully been used to improve genetic medicine in several ways. In particular, large-scale population data has been used to show that variants previously thought to cause disease are in fact common and benign. In some cases the variant has turned out to be more common in the population than the disease they supposedly caused!
In genetics being different is normal!
Perhaps one of the most important changes that needs to happen in genetic medicine is for the baseline expectation to be reset. There is a strong tendency to over-predict the chance that a genetic variant is related to a disease, in part because it is only in recent years that we have realised just how many variants we each have. In turn, this is leading to misdiagnoses and mismanagement in genetic medicine.
When it comes to genetics, being different is normal, and mostly harmless.