NEXTGEN DRUG DISCOVERY ENGINE LEVERAGING ATOMIC-SCALE RNA ENSEMBLES
AI-ENHANCED • STRUCTURE-ENABLED
LEADING THE RNA ENSEMBLE REVOLUTION
Exploiting RNA dynamics
RNAs rapidly exchange between a vast array of distinct 3D conformations, each revealing potentially novel binding surfaces. These dynamics are critical to understanding RNA function and fully exploiting its untapped drug potential.
Our proprietary platform is built on unparalleled advancements in RNA structural biology that allows for the visualization of RNA dynamics at the atomic level with unprecedented sensitivity, enabling the targeting of potentially hidden druggable conformations.
Current Industry Challenges
Much like proteins, RNAs can adopt complex multi-dimensional structures which are involved in a host of cellular functions. Unlike proteins, RNA has its own unique set of challenges that must be addressed.
-
RNA is exceptionally dynamic, a “moving” drug target, which limits the use of conventional structure determination methods and necessitates deep understanding of which conformations will lead to a desired outcome.
-
Selectivity is even more crucial, and must be achieved at the target and even conformational level.
-
Chemical space exceeds 10^63 small molecules, but most physical libraries are much smaller (10^4 - 10^6) and biased towards a few protein classes. A much larger chemical space must be tested to find novel and selective RNA binders with drug-like properties.
-
Current phenotypically discovered hits face significant challenges in development without target structure to elucidate mechanism and guide optimization.
VERITAS PLATFORM
Validated Ensembles of RNA Interactions and Targeted AI-enabled Small Molecules
enhanced by ML and GenAI, grounded in structural biology
5' UTR
3' UTR
CDS
Rapid assessment of new targets for small molecule druggability. Identifying a region in the target that is amenable for small molecule binding and consequential for biologic effect. Looking at structural complexity, uniqueness, energy, and numerous other critical variables.
(example)
Druggable
Region Identification
Structure Guided Lead
Optimization
Once the most promising conformations and compounds are determined, we engage in DART-driven hit-to-lead and lead optimization. We have visibility into the atomic-level pocket interactions to guide specific modifications to improve desired qualities while preserving potency.
.png){kind=small}
The platform then leverages generative AI and combinatorial methods to perform multi-billion-scale compound screens for virtual hits in the matter of days and experimentally validates in NMR, cell free, and cellular assays to confirm interactions and active. Hits are progressed through rapid iterations.
Hit ID
Against
Darts
DART
Determination
DARTs (Dynamic Atomic-scale RNA Targets)
Using a proprietary suite of technologies that leverage powerful NMR observables, machine learning, computational tools and our own internal database, we experimentally determine the atomic-resolution ensemble of the most druggable RNA conformations
SOME KEY ADVANTAGES
DARTs
We leverage orthogonal sources of experimental data and computational methods to distill large and complex RNAs to the key motifs required for desired biological activity, assess its druggability, and generate a dynamic, atomic-resolution, structure of the key element(s). Importantly, these are not models, but experimentally determined and validated ensembles of the RNA target.
Scale & Enrichment
We integrate scalable experimental data with physical methods and ML to evaluate billions of compounds in a matter of days to identify novel chemotypes and match them to their corresponding DART conformations. Enriching hits by over 2 orders of magnitude over conventional screens and enabling targeted medicinal chemistry.
Bespoke Optimization
The Base4 platform is tailored for each target, achieving visibility on the target, region, conformation, and pocket level, critical for specificity and selectivity. This rich insight allows for true SAR. The DART Platform also learns and dramatically improves based on both success and failure, which results in increasingly accurate predictions and faster turn-around times.