How AI is helping to solve the manpower problem in treating rare diseases

Modern biotech has the tools to edit genes and design drugs, yet thousands of rare diseases remain untreated. According to the management of Insilico Medicine and GenEditBio, the missing ingredient for years is finding people who are smart enough to continue the work. AI is becoming the power multiplier to allow scientists to tackle problems that industry has long left untouched, they say.

Speaking this week at Web Summit Qatar, Insilico CEO and co-founder Alex Aliper laid out his company’s mission to develop “pharmaceutical intelligence.” Insilico recently launched its “MMAI Gym” which aims to train major language models, such as ChatGPT and Gemini, to work alongside specialized models.

The goal is to build a multi-dimensional, multi-tasking model, Aliper says, that can solve many different drug discovery tasks simultaneously with superhuman accuracy.

“We really need this technology to increase productivity in our pharmaceutical industry and address the shortage of workers and talent in that space, because there are still thousands of diseases without a cure, without treatment options, and there are thousands of rare diseases that are neglected,” said Aliper in an interview with TechCrunch. “So we need more intelligent systems to deal with this problem.”

The Insilico platform ingests biological, chemical and clinical data to generate hypotheses about disease targets and candidate molecules. Automating steps that once required legions of chemists and biologists, Insilico says it can filter through large design areas, select people who need high-quality treatments, and reuse existing drugs — all at significantly reduced cost and time.

For example, the company recently used its AI models to identify whether existing drugs could be repurposed to treat ALS, a rare neurological disease.

But the labor barrier is not limited to drug availability. Even if AI is able to identify promising targets or treatments, many diseases require intervention at a more fundamental biological level.

GenEditBio is part of the “second wave” of CRISPR gene editing, where the process moves from editing cells outside the body (ex vivo), and to precise delivery inside the body (in vivo). The company’s goal is to make gene editing a one-time injection directly into the affected tissue.

“We’ve developed a proprietary ePDV, or engineered protein delivery vehicle, and it’s a virus-like particle,” GenEditBio founder and CEO Tian Zhu told TechCrunch. “We learn from nature and use AI machine learning methods to mine natural resources and find out which types of bacteria have an affinity for certain types of tissues.”

The ‘natural resource’ Zhu refers to is GenEditBio’s vast library of thousands of unique, non-viral, lipid polymer-free nanoparticles – which are delivery vehicles designed to safely transport gene-editing tools into specific cells.

The company says its NanoGalaxy platform uses AI to analyze data and identify how chemical properties relate to specific tissue targets (such as the eye, liver, or nervous system). The AI ​​then predicts what changes in the delivery vehicle’s chemistry will help it carry the payload without triggering an immune response.

GenEditBio tests its ePDVs in vivo in water labs, and the results are fed back to the AI ​​to refine its prediction accuracy for the next round.

Efficient, tissue-specific delivery is a prerequisite for gene editing in vivo, Zhu said. He says his company’s approach reduces inventory costs and streamlines processes that have historically been difficult to scale.

“It’s like finding a medicine that’s off the shelf [that works] to more patients, making medicine more affordable and accessible to patients around the world,” said Zhu.

His company recently received FDA approval to begin trials of a CRISPR treatment for corneal dystrophy.

Fighting the ongoing data crisis

Like many AI-driven systems, progress in biotech is ultimately running into a data problem. Modeling the conditions at the edges of human biology requires much higher quality data than researchers can currently obtain.

“We still need more real data from patients,” Aliper said. “The data corpus is very biased in western countries, where it is produced. I think we need to have more efforts in the area, to have a limited set of real data, or ground truth data, so that our models can deal with it.”

Aliper said Insilico’s automated labs generate multi-layered biological data from disease samples at scale, without human intervention, which it then feeds into its AI-driven discovery platform.

Zhu says the data needed by AI already exists in the human body, built up over thousands of years of evolution. Only a small fraction of DNA “codes” proteins, while the rest serves as an instruction book for how genes behave. That information has historically been difficult for humans to interpret, but it is increasingly accessible to AI models, including recent efforts like Google DeepMind’s AlphaGenome.

GenEditBio uses a similar approach in the lab, testing thousands of nanoparticles delivered in parallel instead of one at a time. The resulting data sets, which Zhu calls “the gold of AI systems,” are used to train its models and, increasingly, to support collaboration with external partners.

One of the next big efforts, according to Aliper, will be creating digital twins of people to run clinical trials, a process he says is “still in progress.”

“We’re on the plateau of about 50 FDA-approved drugs every year, and we need to see growth,” Aliper said. “Incurable diseases are increasing because we are aging like the people of the world […] My hope is in 10 to 20 years, we will have more treatment options for personalized treatment for patients.”