Specialist
C-level executive at Xtalpi Inc
Agenda
- AI new drug R&D application scenarios and industry development trends
- Major players’ market positioning, main product categories and applications
- Major players’ technological features, data and algorithm support, barriers to entry plus R&D advantages and disadvantages
- Business models, commercialisation process, cooperation models with pharmaceutical companies and cost-saving effects
Questions
1.
Can you first run us through the R&D process for chemical and biological drugs? What are the procedures that AI can play a role in?
2.
What was the penetration of AI technology in different procedures of the drug R&D process over the past 3- 5 years? What are the areas that most AI drug development companies are focused on? There have been trends to apply AI technology in earlier or later stages of the R&D process. What particular trends have you observed? Which areas are easier for AI drug development companies to penetrate or commercialise their products?
3.
According to what you said, AI technology is mainly used for drug screening and design as well as the determining of mechanisms of action. Some industry reports have classified the applications of AI technology in the field of drug development into target identification, compound screening, compound synthesis and polymorph prediction for the pre-clinical stage and clinical trial design, drug repurposing and participant recruitment for the clinical stage. How would you categorise the application scenarios for AI technology in drug development? How big a market is there for each application scenario?
4.
A large number of companies are engaged in the development of AI solutions for target identification and compound screening both internationally and in China. What is the competitive landscape like in the industry currently? How are different companies penetrating the market and how does the application scope of their products vary?
5.
Domestic AI drug development companies include Bioknow, GV20 Oncotherapy, Deep Intelligent Pharma, GigaCeuticals, DeepDrug, ForceClouds and Wuxi NextCODE, as well as some companies focused on polymorph prediction. You said different companies vary in technologies. Which domestic companies do you think enjoy a greater advantage in the market? Which companies are ahead of others in terms of cooperation with pharmaceutical firms?
6.
Could you give us a brief introduction about the mainstream working principles and technological solutions applied in new drug R&D in the mechanisms of action and drug screening sectors? What are the main technologies and how to develop corresponding algorithms or other theories? What are the advantages and disadvantages of different technologies?
7.
What were the technologies and platforms used by R&D institutions and pharmaceutical companies originally? What are the differences between major software for AI drug R&D and traditional software for drug design, such as Discovery Studio, Schrödinger, and Gaussian? What are their advantages and disadvantages?
8.
Comparing with the models developed by major domestic AI new drug companies, what are the advantages and disadvantages of machine learning models? How could AI new drug companies convince the pharmaceutical companies adopting the machine learning models to use their software?
9.
During the entire drug R&D process, there are people suspecting that the application of AI technology will cause difficulties in industrialisation and could affect the experimental results. Is this likely to happen considering that the results are generated manually with the traditional approaches?
10.
In terms of screening molecules with AI technology, the molecules and early stage R&D are actually confidential information for large pharmaceutical companies. How could researchers acquire the data to train AI models? What are the data sources? What is the cost of acquiring data?
11.
How are the computational chemistry methods or machine learning models applied by pharmaceutical companies in developing their AI platforms compared with the system adopted by companies developing new drugs? Are there specific figures to demonstrate the comparison? Will the AI platforms generate a higher success rate or precision than the original models in all cases or just in specific fields?
12.
What are the specific product forms and commercialisation models? Do AI companies simply sell software to pharmaceutical companies or will pharmaceutical companies outsource this sector to AI companies? Are there specific ways in terms of commercial sales?
13.
During the cooperation with pharmaceutical companies or the actual operation, what do you think would be the difficulties and challenges encountered by AI companies? How would AI companies deal with them?
14.
What do you think is restricting the development of AI drug companies? Is it due to the difficulty to convince downstream clients regarding the commercialisation or is it due to the lack of talent? Which aspect will generate bigger challenges?
