Absci Corporation announced the ability to create and validate de novo antibodies in silico (via a computer) with the use of zero-shot generative AI. The ability to create de novo therapeutic antibodies in silico could potentially reduce the time it takes to get new drug leads into the clinic from as much as six years down to just 18-24 months while also increasing their probability of success in the clinic. This new advancement is a major industry step change, unlocking the potential to deliver breakthrough therapeutics at the click of a button, for every patient.

Historically, biologic drug discovery is risky, time-consuming, and expensive, with a >90% failure rate. It takes an average of 10 years and >$1 billion to bring just one new drug to market, limiting the scope and number of treatments that drugmakers can pursue. Absci used zero-shot generative AI — a method that involves designing antibodies to bind to specific targets without using any training data of antibodies known to bind those specific targets.

Absci's model produced antibody designs that were unlike those found in existing antibody databases, and the zero-shot designs worked in the lab right out of the computer — without the slow and costly step of further optimizing the in silico designs in the lab. Absci's breakthrough demonstrates generative AI as an alternative to traditional biologic drug discovery, potentially unlocking treatments for traditionally undruggable diseases and improving therapeutic possibilities for many others. Scalable biological data has been one of the biggest barriers to applying generative AI to biologic drug discovery.

Absci overcomes this challenge with its proprietary high-throughput wet lab technology, which is capable of testing and validating nearly 3 million unique AI-generated designs each week — well above the industry standards. This wet lab data is an invaluable component for improving generative AI models and creating better antibody designs. Absci can accomplish this design to data cycle in a timeframe of weeks.

Absci further demonstrated its wet lab's ability to experimentally validate the superiority of de novo antibody candidates to bind to the target antigen — all without lead optimization of the in silico designs — in cycle times as little as six weeks. Absci validated antibodies for HER2 and multiple additional targets. The achievement is also the first example of a generative AI engine designing new therapeutic antibodies by designing the heavy chain complementarity determining region 3 (HCDR3) from scratch, where the computational design has been wet-lab validated to bind to the intended targets.

HCDR3 is a critical region for antibodies to bind to their targets and enable their therapeutic potential.