Absci Investor Presentation Deck

CASE STUDY: OPTIMIZING HER2 BINDERS Simultaneous co-optimization of affinity and naturalness Predicted Affinity Score 9 1 0 Maximum 5 10 Generation Affinity Objective Minimum Maximum trastuzumab 15 20 Target Maximize, minimize, or tailor binding affinity Naturalness 0.35 0.30 0.25 0.20 0 Maximum 5 10 Generation Affinity Objective Minimum Maximum trastuzumab 15 Target At the same time, ensure sequences appear to come from humans (naturalness) 20 Naturalness 0.35 0.30 0.25 0.20 0.15 0.10 2 Bachas, S., Rakocevic, G. et al., "Antibody optimization enabled by artificial intelligence predictions of binding affinity and naturalness," 2022 pre-print in bioRxiv. ABSCI CORPORATION 2023 ALL RIGHTS RESERVED Best Predicted Variants Training Variants Trastuzumab 4 6 Affinity Score Models simultaneously tuning for affinity & maximizing naturalness 8 absci. 25

View entire presentation