1. Deep Apple relies exclusively on virtual screening for hit identification
According to a recent study, a mere 1% of recent clinical candidates were delivered via virtual screen. 1
Virtual screening represents a huge opportunity because it offers major levers for acceleration, quality, and efficiency.
2. Deep Apple develops its own bespoke virtual libraries
Deep Apple generates project-specific proprietary chemical libraries using our Orchard.ai™ virtual library expansion tool for machine learning-enabled acceleration of hit discovery.
These proprietary libraries are exceeding the diversity of currently available virtual chemical space, enabling better docking against targets of interest, and improving drug-like properties.
3. Deep Apple has a high confidence scoring algorithm for selection and prioritization of compounds coming out of virtual screens
This proprietary scoring tool leverages the attributes of large-scale docking and physics-based methods.
4. Deep Apple uncovers structural biology insights to understand conformations that stabilize or promote specific desired signaling cascades, in GPCRs and other target classes
Our ensemble cryo-EM technology captures dynamics and motion from intermediate structural states. These conformations can represent low population states that may be missed in empirical high-throughput screening, even though these conformations may be the most relevant to biological function.
This approach is designed to uncover new chemotypes, and to do so faster than other approaches.
5. Everything that Deep Apple does is enabled, and continually improved, by deep learning models
Founder expertise in deep learning and neural networks lends Deep Apple unique competitive advantages in the growing field of AI drug discovery.
Orchard.ai™ is Deep Apple’s virtual library expansion tool for machine learning enabled acceleration of hit discovery. Orchard.ai generates project specific proprietary chemical libraries that are > 95% novel against Zinc22, significantly increases dock scoring against the target of interest, and increases druglikeness for hits with high synthetic tractability.
With its proprietary advanced capabilities in structure-based, machine-learning driven drug design, Orchard.ai™ by Deep Apple can rapidly accelerate from target ID to lead optimization in 12 months or less.
