Today is DNA day. Actually everyday should be DNA day. Slower than anticipated – and for good reason – the actual genome revolution, part of the biology revolution is about to start only now.
I recall a TEDx talk I gave on the topic. 10 years ago. You’ll find the link here.
It’s still contemporary. That is not good 🙂
But back to our DNA.
As part of our scouting work to shape the future of health we keep track of novel biological insights linked to one’s genome. Applying this to myself I got triggered by a recent finding regarding specific sleep genes.
I’m a notoriously bad sleeper. Correction, I seem to ‘suffer’ from a condition now named Familial Natural Short Sleep (FNSS). Familial Natural Short Sleepers are genetically wired to have lifelong reduction in nightly sleep duration without evident consequence on cognitive demise, implying that they may have better sleep quality.
All fine. However, numerous studies have demonstrated that poor sleep increases the development of Alzheimer’s Disease. Direct evidence elucidating the benefits of good sleep on the AD pathogenesis is lacking however.
In mice it now has been shown that mutations in 2 FNSS genes lead to resilience to the progression of tau pathology and amyloid plaque formation in neurodegeneration. Hence this specific kind of short sleep protects against Alzheimer’s. A double whammy.
So yes, I looked in detail into my genome. And I’m a ‘mutant’ indeed in one of these FNSS gene products, a transcription factor which regulates other genes. Not at exactly the same site however as described in the paper. But closeby, 15 amino acids from the locus described in the paper, something is really different.
But could that difference have an impact on function?
Hence we checked whether this mutation could have a significant effect on protein function. Therefore we used AlphaFold2, an amazing AI tool from DeepMind on Google Colab, or its runner-up RoseTTA Fold. This tool stirred up the community when it turned out to be able to accurately predict protein 3D structures, simple by calculation. No more difficult experiments making crystals required. Simply copy-paste the protein sequence into your browser and press “run”. That’s how easy it is today. Amazing to have such a powerful tool at our fingertips and kudos to the teams that make these tools freely available to all, helping advance science.
Both the mutation reported in the paper and ‘my mutation’ are in a region that is required for the binding to other proteins or DNA. Where in the wildtype protein there are two consecutive alanine amino acids, my mutation shows two consecutive prolines (in red on the figure). This is significant as it is known that proline acts as a structural disruptor in the middle of regular secondary structure elements such as alpha helices and beta sheets. So having two prolines in a row for sure must have an effect on the structure or binding properties and by consequence on protein function in my body.
So what does this tell us? Well wait, not so fast. Did I mention a risk for Alzheimer’s disease? Is that something you would like to know ? And could you do something about it if you knew?
I do remember 10 years ago when I analysed my own genome, being ready to reveal the mutation status of the single gene (ApoE4) at that time linked to a significantly increased link to develop Alzheimer’s disease.
My heart was beating 250 beats per minute I believe. But mentally, I had prepared for when the news would be ‘bad’, I would cope. By starting some brain training. The topic for another post.
Since April 4, we know a large number of novel Alzheimer’s disease predisposition genes. Hence it becomes possible to develop an Alzheimer risk score.
But do people want that? Or how can we distinguish people who even don’t want to know you could know this, from the others?
That requires novel healthcare jobs. Alfa, not Beta. Not just biomedical experts and data scientists, but empathic human genetic counsellors and educators.
But also ways to translate someones wish (I want to know, anyway – I only want to know if I can do something about it – I don’t want to know) into a personal data store, a POD. Like what We Are is developing. The Belgian citizen-centric data platform, officially to launch June 2022.
So what if my POD contains my predisposition for Alzheimer’s disease as well as for say eye-diseases (that’s my biggest fear, to get blind before my brain stops behaving).
Which services would I like to receive? Yours? Offering #digital health tools to assess brain and eye health at home? Offering diet advice on how to delay disease? Offering brain training ?
Hence, to sum up what is needed to make everyday DNA day:
- Biomedical progress
- Lots of novel biomedical insights, with much smaller gaps between an initial discovery (often in a good journal) and actual translation and validation in applicable medical practice. In the case above it will start with novel experiments to confirm these (mice) data in humans.
- Large number of people getting sequenced, taking ethnicity into account.
- Automated tools to translate putative DNA mutations in 3D models allowing drug discovery teams at pharma & life sciences companies to apply this knowledge
- Novel jobs
- To deal with privacy (and thus cybersecurity) aspects
- To translate biomedical insights to healthcare professionals first, patients and citizens next (education / learning)
- To assess the ethical aspects, even the moral aspects of these findings and how to deal with them
- Translation into health advice. For the price of a mobile phone, you can now have your genome sequenced and obtain valuable and actionable lifestyle and health advice. It can no can no longer merely be ‘sleep more’. We need context.
Excited as we are? In our keynote or workshop (one artistic version entirely by way of cartoons) we show the art of the possible today and we welcome you to start exploring anything genomics yourself. More info to be found here.
BTW, I fully realise that I paint a simple picture here. I know DNA is only one layer of our omics composition. I know some mutations do not translate in protein to be modelled, but rather in long non coding RNAs. The list is long. But I hope you see my point. The day is called DNA day.