How Himalayan communities thrive in breathless heights

And can AI create the next antibiotic?

Hello dear, friends!

It’s been brutal out there. I hope you’ve been able to stay cool. I’m off to India in a few days (New Delhi and Kolkata for sure, with possibly an added trip somewhere else).

For my main topic this week, I’m going to turn to three things I care about deeply - microbes, antibiotics, and AI.

Be sure to read something in-depth about this from me when the time is right (hint, hint), but I do believe that antimicrobial resistance is one of the unheralded challenges of our times. COVID-19 was bad. Now, think worse.

Alarmingly, this menace has already infiltrated the real world. The World Health Organization reports that this is among the top 10 global public health threats facing humanity.

The pipeline for antibiotics is failing to keep pace with the burgeoning challenge of antibiotic resistant superbugs.

Worse, the market for novel antibiotics is virtually nonexistent. The costs of development, manufacturing, and distribution dwarf the return on investment for new ‘reserve’ antibiotics.

Can AI make antibiotic discovery faster, cheaper, and easier?

That’s the topic of my science column this week which you can read here.

The short answer is that it’s still too early to know. If LinkedIn reposts were a reflection of what happens in the real world then AI would’ve solved every problem facing humanity by now, including superbugs.

Here are the key points of my column:

• Drug-resistant microbes killed more people worldwide than malaria or HIV in 2019.

• AI might hold the key to battling these superbugs, with research showing promise in drug discovery, specifically for antibiotic development.

• A research team has used deep learning to identify promising antibiotic compounds, two of which - halicin and abaucin - have shown potential in fighting specific superbugs in lab settings.

• The AI model applied to drug discovery works by learning from vast libraries of chemical compounds, which could potentially speed up the drug discovery process and reveal treatments humans may overlook.

• A specific example of AI's potential is seen in its battle against a deadly superbug, Acinetobacter baumannii. The model was trained to recognize compounds that halt bacterial growth, leading to the identification of abaucin.

• Abaucin's strength lies in its potent action against A. baumannii and limited effect on other bacteria, preventing the harmful consequence of killing good gut bacteria, a common issue with broad-spectrum antibiotics.

• AI has also been successful in identifying halicin, another potential antibiotic effective against E. coli, proving the method's replicability and potential.

• The research team plans to extend the AI-based approach to other superbugs, aiming to expand the range of antibiotic effectiveness.

• AI's potential isn't limited to new drug discovery; it can also optimize existing antibiotics and help with disease diagnosis and treatment, even in non-infectious diseases like cancer.

• While there are exciting possibilities, AI hasn't led to a single approved drug yet, highlighting the long and complex path from drug discovery to market.

So like most things- reality is more complex than viral video short by The World Economic Forum.

---

As a plains-dwelling lowlander I suffer whenever I’m in the heights of the Himalayas and Andes.

But populations like the Sherpas have thrived in such conditions for millennia. This is thanks to specific genetic adaptations that improve oxygen efficiency.

In response to hypoxia (low-oxygen levels), lowlanders' bodies increase ventilation, cardiac output, and produce more red blood cells, whereas Sherpas and many Tibetans avoid this blood-thickening process. How do they do it?

Sherpas use oxygen more efficiently by burning glucose over body fat, due to an advantageous genetic mutation.

Some Tibetans and Sherpas have a unique genetic adaptation, a gene called EPAS1, which they inherited from an ancient hominin lineage, the Denisovans. These are an enigmatic human ancestor that has only been recently discovered.

The Denisovans disappeared around 50,000 years ago, but their genetic legacy, particularly the EPAS1 gene, helps Tibetans and Sherpas survive high altitudes without the risk of mountain sickness and heart attacks.

Denisovans also passed on DNA to modern human populations across Russia, eastern Asia, and some Pacific islands.

Andeans, another high-altitude population, developed unique traits like an enlarged chest and increased lung capacities to survive the low oxygen conditions. Genes involved in cardiac function, oxygen sensing, vasodilation, and oxidative stress reduction show signs of positive selection in Andeans, enabling their high-altitude survival.

So, there isn’t one way to thrive in high-altitudes. But some communities have adapted to these harsh conditions for hundreds of years!

That’s it for now. Stay cool and hydrated!

Anirban