- XploreSeqTM analyzes millions of antibody sequences and recommends a highly-diverse panel of candidates with >50% hit rate
XtalPi is a global pioneer in AI drug discovery. Founded in 2014, we have partnered with 16 of the top 20 pharmaceutical companies in the world.
XupremAbTM is our one-stop solution for antibody discovery that incorporates AI throughout the whole process. The platform can be accessed as a whole, where we can take a partner’s project all the way from target to preclinical candidates, or it can be accessed as individual sub-platforms.
Development of XupremAbTM was guided by the following principles
Combine the best of AI and wet lab
AI is a powerful technology, and its impact is maximized when combined with wet lab. We have built a fully-functional antibody lab with a cross-disciplinary team of AI scientists and experimental biologists. This synergy creates new opportunities that cannot be achieved through wet lab or dry lab techniques alone.
Tap into greater search space for more quality hits
To lay a solid foundation for any discovery campaign, it’s essential to have as many diverse, high-quality hits as possible at the start. We use AI to enhance existing experimental techniques or let AI take the lead to expand our search space for hits.
Utilize AI to make antibody engineering more efficient, enabling multi-objective optimization
We want to reduce the guesswork and random mutagenesis in antibody engineering. AI helps us explore a larger mutant space and account for multiple properties simultaneously when selecting candidates.
Instill excellent developability into every molecule
Developability is an essential property that we capture in the discovery phase. A combination of in vitro and in silico Instill excellent developability into every moleculein vitroin silicomethods allows us to assess developability at ultra high-throughput and reasonable cost.
- Tap into greater search space for more quality hitsCase 1In silico screening of millions of antibody sequences in the immune repertoire
XploreSeqTM routinely sequences millions of BCRs and identifies hits with high confidence (left).
In this project, XploreSeqTMpredicted 717 binders, from which 48 were randomly chosen for expression and testing. 66.7% (32/48) were confirmed to be binders, and only 18.6% (6/32) binders overlap with hybridoma hits (right).
Utilize AI to make antibody engineering more efficient, enabling multi-objectiveoptimizationCase 2Bi-directional affinity tuning of a lead antibodyGoal: tuning the affinity of a lead antibody both upwards and downwards to provide more options for bispecifics developmentOur AI model generated 3 buckets of designed mutants with predicted affinities of:
1) significant increase (target: 10x),
2) mild increase (target: 3x) and
3) mild decrease (target: -3x) vs. WTWhen expressed and tested, these mutants show affinities that correspond to their targeted ranges. (Shown on the right)1 round of design-make-test of 96 mutants to arrive at a panel of highly optimized leads with KD that span 3 orders of magnitude and optimal developability profiles.
Instill excellent developability into every moleculeCase 3XcelDevTM Silico evaluates and ranks 30+ candidates in terms of developability36 hits that had passed binding and functional screening were analyzed by XcelDevTM Silico. 6 properties of each antibody was predicted and scored on a scale of 0-1 (displayed in color gradient).
An overall developability score was calculated to rank all 36 hits. To verify the effectiveness of the ranking, the top 7 (in the red box) were expressed and subject to a battery of developability assays. All 7 performed well in these assays, their Tm, SEC, AC-SINS and HIC results shown on the right.
XtalPi is keenly aware that every research partner has unique needs and priorities, and is happy to create a customized solution to optimize your research program
