Computational Aging

Are we making progress in life extension? Will it be in my lifetime?

Here's a signal that we're making real progress in life extension: when we have interventions that allow mice, rats, dogs and monkeys to live much longer and healthier than they currently do.

We're not there yet.

We've seen some incremental evidence of life extension in mice. Still, we can't get a mouse to live reliably to 5 years old and beyond. We can't get a rat to live much beyond 6 or 7, though maybe Harold Katcher's blood factors will change that?

Mice and rats are interesting, especially given the naked mole rat. Naked mole rats can live for 30+ years. They age much more slowly. They have better DNA repair. They get less cancer. They're similar to rats and mice (though more similar to guinea pigs, which live 5-7 years).

One interesting research path would be to make mice and rats (or guinea pigs) more like naked mole rats. Maybe gene editing or drugs to stimulate the same DNA repair pathways?

We also haven't seen much life extension yet in dogs and monkeys. The Dog Aging Project may change this soon. This study is giving rapamycin to older companion dogs.

Epigenetic reprogramming is promising and high-visibility. We'll know it's especially promising when we see mice, rats, dogs and monkeys living longer and healthier because of it. Same with the factors in young blood.

Put another way, if we can't reliably increase lifespan and healthspan in mice, rats, dog and monkeys, we probably not that close to achieving it in humans.

One of the most important questions within aging science is identifying good biomarkers of aging.

We need biomarkers that are predictive of aging rate and mortality. We have lots of biomarkers that are associated with aging rate and mortality, like cholesterol readings and epigenetic clocks. But we don't know how predictive they are.

We need these biomarkers to be more frequently measured than a full human lifespan. That way we can test interventions and see how they work. The more frequently these biomarkers can be taken (and still be predictive), the more we can make lots of personalized interventions (e.g. lifestyle, drugs, epigenetic reprogramming, gene editing).

Once we have these biomarkers, I could imagine a device like an Apple Watch giving us personalized recommendations based on population and personal data on how to slow our aging. Apple Watch captures large amounts of physiological and lifestyle data. They are increasing their data-capturing capabilities every year.

Working on biomarkers of aging is likely a very good use of time.

“If I can dream, in a few years we’ll have spatiotemporal single-cell ’omics in living tissues.” – Sten Linnarsson, Karolinska Institute[1]

When we know exactly what is going in every tissue and cell of the live human body, we probably will figure out how to cure aging.

But how much data storage is necessary to track every protein and non-coding RNA in the human body?

Here's a quick estimate:

• 30 trillion cells per human
• 80,000 unique non-coding RNA
• 20,000 unique proteins
• = 5x10^22 measurements per human
• Now, measure this hourly every year (8,760 hrs/yr)
• = 5x10^26 measurements per human per year
• Call it 4 bytes per number (float32)
• = 2x10^27 bytes per human per year
• Convert this into zetabytes at 10^21 bytes per zetabyte
• = 2x10^6 zetabytes per year per human
• Measure this across 10 billion humans
• = 2x10^16 zetabytes for humanity per year

So we need 2 million zetabytes (2x10^6) per human per year and 10 billion times more for every human. Unfortunately, we only captured about 100 zetabytes of data as a species in 2021[2]. This math is probably wrong... but probably at least directionally right.

We don't have the ability to store (or of course yet capture) this amount of data. But single-cell spatial transcriptomic methods will continue to rapidly improve, as will data storage and data science methods. For example, we may get tissue-level live 'omics in the next few decades.

[1] Marx, V. Method of the Year: spatially resolved transcriptomics. Nat Methods 18, 9–14 (2021). https://doi.org/10.1038/s41592-020-01033-y

[2] "Volume of data/information created, captured, copied, and consumed worldwide from 2010 to 2025." Statistia.com. Accessed June 24, 2022. https://www.statista.com/statistics/871513/worldwide-data-created/