Drug development productivity has always been about balancing multiple variables to get most out of your R&D dollars.
Success rates, cycle time, costs, number of experiments, value of outcomes. This is where GenAI for proteins is different from previous innovations. What we do at Cradle has direct impact on all of these:
🎯 Success Rates (p(TS))
Probability of approval through clinical trials increases as candidate quality increases (higher potency, lower tox/immunogenicity, easier delivery). You may stills stop perusing a molecule for commercial or translation reasons, but not hitting your target product profile is much less of an issue.
🔄 Cycle Time (CT)
The number of cycles required to get through hit-identification (i.e. screening) to find the best lead candidates as well as the number of cycles required for lead-optimisation come down. And if you invest in automation or the right CRO you can significantly reduce experimental cycle times for many assays from 8-12 weeks down to 2-3 weeks..
⚡ Experimental Throughput (WIP)
On top of requiring fewer cycles, GenAI will allow you to only run experiments that actually need to be run.
1) avoid running assays (i.e. predicting t-cell activation throughout your project and only having to run a confirmatory assay),
2) be more sample efficient (i.e. if models are confident you can run fewer samples to save money)
💰 Costs (C)
We don't really impact cost of labour, re-agents and equipment. But luckily, cost of sequencing, DNA synthesis and lab equipment are all on fast downward slopes. This tailwind is there for everybody.
🥇 Value to the organization (V)
GenAI (in particular de-novo design) can be used to make products against hard-to-target targets allowing you to build a lot more differentiated products.
These different factors are multiplicative and I expect we will see strong R&D productivity growth over the next few years.
It does all start all starts with putting the right tools into the hands of scientists. Send us a message 😉.
