The Economist March 14th, 2020 pp4-8 Technology Quarterly|Personalised medicine
The Human Genome Project “Welcome to you”. “ Twenty years on, genomics is really starting to matter”
Mapping the human genome was a necessary but not sufficient step in widely leveraging molecular genetics for medical intervention. Before the human genome project (HGP), many obvious genetic diseases had been identified and their molecular nature elucidated. These early examples were by today’s standards relatively simple because a clearly defined mutation resulted in obvious disease. For those disorders the patient manifests disease at birth or later in life. Although these mutations are rare definitively knowing the patient's gene status is absolutely predictive.
HGP a critical foundation for beginning to understand the more complicated molecular mechanisms causing disease. Much more intensive study is now being pursued to better understand diseases involving multiple mutations, involving the regulation of gene expression and those involving many genes operating in combination (Polygenic Disease).
As an example of a polygenic disease, type-2 diabetes is now known to be associated with at least 94 genes. “Some of these genes have variants that increase a person’s risk of the disease, others have variants that lower risk”. None of the variants contribute much risk but carrying many high-risk variants is significant. Enter the “Polygenic Risk Score” (PRS). The higher your PRS the more likely you are to develop disease. For a form of breast cancer, 313 variants have been identified, those with the highest PRS “are four times as likely to develop cancer than the average”.
These scores are not based on gene sequencing, rather they are based on genetic variants known as single nucleotide polymorphisms or SNP. A SNP, at any particular base pair location, on the genome can be evaluated on a population basis by sequencing many individual genomes. As an example, at a SNP site most people have a specific nucleotide base-let’s say C, for cytosine, but a small but significant fraction of people have-let’s say A, for adenine. Unlike a rare mutation, in this example, A for C is present in more than 1% of the population. As used in human disease association, it may be that having an A for C at a defined base is associated with a specific disease. Important SNPs, those conferring some disease risk, are thought to be near genes actually involved in disease expression.
Now, compared to gene sequencing, interrogating an individual’s genome for SNP sites is relatively quick and inexpensive. “Over 70,000…[SNPs]…have now been associated with diseases in one way or another”. “Some argue that these scores are now reliable enough to bring into the clinic". Using these data clinicians could help patients by customizing approaches to further disease screening, help with efforts to quit smoking, help with treatment of behavioral disorders and help select the most effective pharmaceutical treatments.
Some worry about the inappropriate use of SNP and sequencing technology. As we are all aware, there has been an interest and demand for "direct to consumer" genetic testing. Companies like 23andMe generate “information about your predisposition to diabetes, macular degeneration and various other aliments”. These data from "direct to consumer" testing are interesting but not actionable.
Many ambitious sequencing projects are underway to further catalog human genomes. Over time the speed and cost of the whole genome sequencing (WGS) has fallen dramatically. “Today various companies…[offer WGS] for $600-$700 and companies are targeting a $100 price point. “As yet, there is no compelling reason for most individuals to have their genome sequenced. If genetic insights are required those which can be gleaned from SNP-based tests are sufficient for most purposes”.
Lastly, “Genes are not everything. Controls on their expression-epigenetics, in the jargon-and the effects of the environment need to be considered too…”