Future of Healthcare - Genomic Medicine: Part 1
As a PhD student working on Finnish population genomics I am quite fascinated by how researchers have begun to integrate population genomics and modern medicine. In a series of articles, I would like to share my views on the future of healthcare i.e., genomic medicine.
A wee bit of history!
The past two decades have revolutionised medicine at an impressive rate with genomic, post-genomic and modern bioinformatics approaches contributing heavily. It all started of course with the extraordinary, ground-breaking Human Genome Project, taking 13 years and 2+ billion euros for completion which ended up sequencing 3 billion base pairs with the goal of mapping the human genome. Major proposed applications for the project included looking for genomic variants which alter the risk for certain diseases or cataloguing mutations within cancer cells. The project indeed was a grand success paving way for functional, comparative genomics, better bioinformatics and sequencing methods. For example, price of whole genome sequencing has reached 1600€ and the day is not far when pocket next generation sequencing (NGS) would be available for everyone. I am sure many of us are aware of the figure by the National Human Genome Research Institute given below:
The entire NGS market is expected to reach 20 billion $ by 2020 and many private companies like Illumina, Roche are vying for a giant slice of that pie.
However, HGP brought major surprises–fewer genes than expected (20k than the 80-100k believed before), existence of large swathes of non-coding DNA, large amounts of repetitive DNA, bunch of genes not found in invertebrates and problematically, it didn’t directly help us cure major diseases. This led geneticists to conclude that DNA is not the lone gun but several “other” players exist (more on this later).
In parallel to HGP, several international projects like ENCODE and HapMap were also launched. We now have quite a robust collection of protein structure, sequence, RNA, genome, gene expression, metabolomic databases which all are part of the effort to find more about the “other” players. In addition, large-scale genome wide association studies (GWAS) started off to hunt the disease-specific casual genomic variants. And hence started the exciting post-genomic era of figuring out the puzzle of which genetic components combine together to bring about the complexity of human physiology and disease.
Though causal genes for a lot of Mendelian disorders were discovered, many large-scale exome-sequencing studies which focused on protein-coding regions alone, found about 40% of the genetic basis of complex diseases. However, previous estimates from genetics had estimated heritabilities greater than 70%-80% for various traits like hair curliness, schizophrenia, IQ and height. If the genetic variants have really high cumulative effects - where were they? The missing heritability issue is one of the most widely discussed and debated topics among scientists for the past decade, as the initial promises and high expectations from HGP and other sequencing projects have not yet matched the actualities of complex disease genetics.
Some reasoned that the sample sizes for such GWAS were quite low, and in accordance with that, studies like a recent one on blood pressure which involved samples of >1 million came into existence. Others recommended discarding the common variant-common disease hypothesis, in favour of rare variant-common disease, which emphasizes the role of rare variants with large effect sizes. Hence, people have reasoned to have larger GWAS or to go back to the early 80’s-90’s style family based genetic studies (trios/pedigrees).
So, is the solution simply, to increase sample size and do bigger GWAS? Would non-coding regions, splice sites, common variants etc. taken together solve the heritability issue? And even if we do find the causal variants can we explain, how do these genetic variants (some causal, others less so) raise the risks of complex diseases? Would we be able to connect genes & regulatory regions to phenotypes via biological pathways, which can then be used for biomarker development for diagnostic tests and for drug targeting?
Understanding which genomic variants are to be detected and their interpretation, along with detailed functional studies of molecular mechanisms is an important next step for researchers worldwide. For only then, we can move towards a better understanding of the biological context of disease-specific genomic variation for more effective diagnoses and designing treatment techniques.
Answering these questions form the core for the next few decades of medical science & genomics and would truly usher us into the era of personalized healthcare, the holy grail of modern medicine.