Opinion
When I was young, I broke my arm playing football. The coach decided I was fine, but when I got home my mother took me to hospital just in case. In the emergency room a doctor asked me – how great is the pain, on a scale of 1 to 10?
I said 7. I didn’t want to minimize my pain, but nor did I want to exaggerate. I remember very clearly having no idea whether 7 was accurate. I just thought it would suffice as an answer.
Rendering qualitative things into quantities, even just into discrete categories, can be very useful in science. Once things become numbers (and ideally when they become integers, i.e. exact numbers) they can be more easily stored and analysed. In the modern age many numbers can be stored, and then subjected to computational analysis, and most recently to analysis by AI.
I also remember when the university started keeping recordings of lectures. A staff member came to me with a plan to digitise the recordings – to convert all the spoken words into written text. I asked – why do we need to do that?
He looked at me patiently (but in disbelief), and explained that once you render a qualitative thing, an analogue soundwave, into something discrete like letters, then you have a digital record that can be searched by computer. Students could type in key words and search entire lectures series for topics they wished to revise.
Over the years I’ve seen many things, museum and library collections, lecture recordings, music, and broadcast media, being ‘digitized’.
I’m still often shocked by how imprecisely qualitative ‘waves’ are converted into discrete categories – I refuse to believe there are precisely seven colours in the rainbow – but I certainly recognize the value of assigning numbers and thus allowing a level of precision.
During my career the importance of DNA and genomics has been increasing, and it occurs to me that nature chose DNA as the genetic storage material because it, and its cousin RNA, are both inherently digital. They are already precise.
DNA has sequences made up of 4 chemicals (named A, C, G, and T). They are like exact integers. There are no half measures. If there is a change, it is a complete change, a mutation is, by definition, one letter being swapped for another.
This absolute precision means that DNA data is immediately amenable to computational analysis. I expect that AI will have a ball with human genomes in the future.
When it comes to genomic research, I believe the possibilities for Precision and Computational Medicine will be very great. Interest in this area will accelerate rapidly as most, if not all people, have their genomes sequenced.
But why do I think everyone will get their genomes sequenced?
Some individuals will opt in out of interest, perhaps to understand their ancestry. Others for health reasons. About 10% of families harbour inherited diseases and many will want diagnoses. But even more people will get cancer samples sequenced and compared to their normal non-tumour genome, to identify mutations that turn on cell proliferation pathways that can be targeted by specific drugs to reduce cancer growth.
Since families share DNA, once one member of your family is sequenced, light will be shed on your own genome (whether you want to know the secrets of your genome or not!).
Genome sequencing may soon become the norm. In the UK there is a plan to offer DNA sequencing to every newborn child.
Currently this is about diagnosis. But importantly, our ability to not just read, but also to re-write genomes with CRISPR editing, is rapidly progressing.
As more harmful mutations are discovered, the pressure to understand DNA-based conditions and devise new genetic treatments will increase. There will be growing interest from patient groups, health professionals, scientists, governments, industry, and philanthropists.
Experts at universities will have a role to play. My own university, UNSW Sydney, has long emphasized quantitative research (the ‘hard sciences’) and has huge schools of Computer Science & Engineering, and of Mathematics & Statistics.
Fortuitously, over the last few years the land between our campus and the vast Randwick Hospital precinct has been resumed and now consists of jointly shared buildings physically linking the university to the health sector.
Not surprisingly our strategic plan includes developing a focus on Precision and Computational Medicine.
Gradually biology and medicine have moved from the study of analogue blobs, to the precise analysis of DNA sequence data, which is already digital and readily computable.
At present, we know a lot about ‘single gene disorders’ but increasingly we’ll be tackling diseases that arise from multiple different mutations, and we’ll want to understand statistical probabilities of occurrence and the efficacy of treatments. DNA is digital but the outcomes, our physical form and our brains, and ultimately human culture, will always exist as rich analogue outputs, so there will be work for everyone, not just data scientists.
Biology has transitioned through chemistry, to now be dominated by the data science of DNA. Just as weather forecasting began very slowly before becoming accurate, the data science of genomic predictions is just beginning, but already it is clear that the vast, growing reserves of digital data on human genomes will be hugely powerful.
I predict great opportunities but also great risks as this field develops, and I am looking forward to seeing the age of genomics and the eras of both diagnosis and treatment via CRISPR and RNA therapies, unfold.
Professor Merlin Crossley is DVC (Academic Quality) at UNSW