Absci Investor Presentation Deck
CASE STUDY: AI-DRIVEN LEAD OPTIMIZATION
85% of Top 100 "natural" Trastuzumab variants
exhibit higher-affinity than wild-type
Count
20.0
17.5
15.0
12.5
10.0
7.5
5.0
2.5
0.0
Designed Variants
Trastuzumab
7.6
8.2 8.4 8.6 8.8
-log10 (KD)
85% of top binders have higher affinity than Trastuzumab
7.8
8.0
9.0
■ Al predicts the affinity of unseen variants from
libraries generated using diverse mutational
strategies and combinatorial sequence space
Al models make predictions with actionable
performance using <0.1% of the combinatorial
sequence space as training set
■ Naturalness is associated with developability
metrics and expression titer
▪ Enables one-shot multiparametric lead
optimization potentially accelerating time to clinic
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
absci. 26View entire presentation