Software is eating drug discovery. Welcome Qubit Pharma!

Qubit Pharmaceuticals is accelerating drug discovery thanks to quantum chemistry & physics  powered by  hybrid high performance computing & quantum computing platforms. Their goal is to cut cost and length of preclinical drug development with their simulation platform, Atlas. Thanks to its unprecedented precision, Atlas allows the generation of digital twins of molecules, a breakthrough that allows drug discovery to move from approximation to prediction.

The company has extracted the scientific work of five French and American scientists on the development of polarisable force fields massively accelerated by latest generation molecular dynamics software. At XAnge, we are strong believers in  this sector, at the frontier between biotech and tech, called  the CADD (Computer Aided Drug Discovery) market. This market is experiencing a take-off in terms of investments, having raised from $2 billion in 2015 to $16 billion in 2021.

Here are the most significant pharma deals driving this market :

• A potential $2.7bn deal including $55m upfront between Bristol Myers Squibb and Schrödinger.
• A potential $1.2bn deal including $50m upfront between Bristol Myers Squibb and Exscientia.
• A potential $12bn deal including $150m upfront between Roche and Recursion.
• A  potential  $5.2bn  deal  including  $100m  upfront  and  $350m  per  molecule  found between Sanofi and Exscientia.

Those milestones participate in the massive digital boom in drug discovery, industrializing the generation of novel drug candidates.

Why do the costs of drug discovery drive pharmaceuticals industries into CADD?

Computer-aided drug design (CADD) includes several techniques covering finding, designingand analysing drug candidates and their related biological targets by computer methods. The use of these methods speeds up the early stages of novel drug development while supporting and speeding up drug discovery.

Although the estimates are highly controversial, today, less than 5% of drug discovery projects initiated reach the market, for a total development cost per successful project of about $2.6 billion per new drug.

For over six decades, computational power has grown exponentially under the rule of Moore’s law. Most industries received enormous productivity boosts thanks to this tailwind. The pharmaceutical industry is a notable exception. The cost of R&D to bring a new drug to FDA approval has been doubling roughly every nine years—in what came to be known as Eroom’s law

Thus, while we got faster and better at making cars and phones, it became harder to discover new life-saving treatments. Biotech & pharma spend about €2bn for each drug that reaches the market. About 30% of this budget (€570m) goes into generating lead candidates. This part of the process (known as drug discovery) takes 3.5 yrs on average, and each lead has just a 12% chance of success. The consequences are clear: expensive drugs, risk-averse discovery programs, and decade-long development cycles

Drug discovery is very  expensive for these main reasons:

• First, each new drug must be better than the previous available treatment, raising the level  and making the technologies more and more complex.
• Secondly, most of the new drugs only have modest incremental benefit over the drugs already considered as successful. The smaller increment of these treatment effects calls for an increase in clinical trial sizes to show the same level of efficacy.
• Third, despite real scientific advances, we are just beginning to have sufficient biological knowledge and computational power to build robust models to understand how a protein will respond at the protein, cell, organ, and whole-body levels, resulting in a high failure rate in the process.

We are at an exciting time in History when science and computing power are coming of age. These two forces will continue improving, making these new technologies more and more striking and giving hope for a new wave of strong innovation in biotechnology. 

Qubit makes drug discovery more efficient

Let’s recall the drug discovery process: it starts by studying the biology of a cellular target. Once the biology is well characterized, the process continues with high-throughput screening of thousands of existing chemical compounds to find those that ‘hit’ the target of interest.  This approach is slow, constrained by the size of the physical chemical libraries and subject to high attrition rates. As a result, only one in 10,000 molecules sought for a given target will be found.  

The technology developed by Qubit Pharmaceuticals  allows to obtain a great speed and cost optimization in these steps by being full in silico. Qubit is of course not the only start-up to take up the challenge! But they have developed particularly powerful engines and models allowing them to be fast and reliable from the target characterization up to the lead optimization phase. Moreover, existing simulation software can mostly only handle 30% of  the total value of the drug market : because of the lack of precision of their physics models or of the duration (and thus costs) of calculations, tools had to be specifically tailored to certain types of targets such as kinases and enzymes.

To explain briefly Qubit’s methodology in four steps:

• Develop dozens of numerical twins of targets of interest with great accuracy, allowing to structure and aggregate a considerable mass of knowledge.
• Build their own supercomputer (25 pflops to start). Combine it with the computing capacities available on the cloud in off-peak periods in order to have a very high capacity computing power at low cost.
• Use the specific technology developed by the company to screen drug candidates by being able to accurately predict the binding strength between the drug candidate and the target of interest.
• Push the best drug candidates into the clinical phases. Either internally or through partnerships.

In 2020, Qubit raised a pre-seed round that allowed the technology to be transfered from the laboratories and to structure a high-level core team. Today we are happy to join the adventure to meet the challenges of Qubit in this shinny seed round!

As a new board member I am thrilled to partner with the scientists’ founders and the management, working on new computational approaches to design faster medicines that are more affordable.