FB Pixel no scriptInsilico Medicine to test AI-discovered drug candidate on humans this year | KrASIA

Insilico Medicine to test AI-discovered drug candidate on humans this year

Written by Song Jingli Published on   3 mins read

Insilico has proven that AI can solve very difficult problems in a creative way, says Sinovation Ventures’ Kai-Fu Lee.

Hong Kong-based biotechnology company Insilico Medicine is expected to start clinical trials before the end of the year for a drug that was discovered by its AI system.

“We are designing an AI-powered clinical trial now,” Alex Zhavoronkov, founder and CEO of the company, told KrASIA on Wednesday.

The firm announced last week that it developed the drug candidate which is treating idiopathic pulmonary fibrosis (IPF), a rare lung disease, and other fibrotic disorders, within 18 months at a cost of USD 2.6 million in total. The process normally takes 2.5 to 4.5 years and costs about tens of millions of US dollars, said chief science officer Ren Feng during a webinar on Tuesday that was co-hosted by Insilico and its investors Sinovation Ventures and Qiming Venture Partners.

Zhavoronkov founded Insilico in 2014 when deep learning systems started to outperform humans in image recognition and a technology called generative adversarial networks (GAN) was invented. “To my knowledge this is the first case where AI identified a novel target and designed a preclinical candidate for a very broad disease indication,” Zhavoronkov said last month.

uptake newsletter

In a traditional research process, human scientists identify a target, usually a protein or gene that partly causes the disease, and then discover or syndicate a chemical compound to hit the target.

“We developed more than 60 different AIs and advanced bioinformatic engines for target discovery,” Zhavoronkov said on Thursday. PandaOmics is one of the systems that integrates many of these scoring engines, he explained. It has many scores that use human “omics” data to identify novel targets that were not previously linked to a disease, and another set of AIs trained on text data to evaluate the level of novelty and confidence looking at the indirect links implicating the target into a disease.

“One of the absolutely unique approaches developed by Insilico is a deep learning system utilizing omics data to predict human biological age,” he said. The goal is to understand the small changes that transpire in tissues and cells during normal aging in the absence of other diseases. Insilico then uses these networks to understand what targets may be defensive against aging, to be able to launch the body’s natural defenses, and also what targets are potentially harmful and should be disabled. This new approach, he added, helps distinguish the most important targets in diseases from very small amounts of available patient data.

Imagining the future

Generative biology and generative medicine are other approaches the company is working with, said the CEO, who is also a PhD in physics and master of science in biotechnology. Insilico uses GANs and other generative AI approaches to “imagine many versions of just one patient or a group in the future by creating millions of their copies with a different age and to imagine how changes of one or more targets at the starting ‘template’ patient can impact the future versions of that same patient,” he added.

To inhibit the specific target discovered for IPF, the company’s Chemistry42 system, powered by Nvidia V100 Tensor Core GPUs, that uses a structure-based drug design generative-chemistry approach, also came up with a set of novel compounds, which are bioavailable, metabolically stable, capable of oral administration and safe, according to the company’s press release last month.

“Insilico Medicine has integrated AI in the entire new drug discovery R&D process and proven AI can solve very difficult problems in a creative way,” Kai-Fu Lee, founder of Sinovation Ventures, said during the webinar.

Nisa Leung, managing partner of Qiming Venture Partners, explained that in the medical care sector, it takes increasingly more time and money to get new drugs, and that the success rate has been declining. New technologies are needed to improve the success rate of new drug discovery, to improve development efficiency and reduce costs, she added.

Her firm started to invest in the digital medical sector in China as early as 2010, backing a slew of companies that have grown to be public companies including Gan & Lee and Zai Lab.


Auto loading next article...