Drug design has typically focused on optimizing molecules for activity and, in most cases, a few simple ADME properties. While this process made some sense in the early days of in silico ADME, it does not consider what is ultimately important: will the molecule(s) you design be active to the target, have minimal toxicity risk, AND meet preclinical/ clinical PK requirements?
The ADMET Predictor AIDD Module combines technologies in ways in which no other company has done. The combination of machine learning and GastroPlus® PBPK modeling, wrapped within proven generative chemistry methods, allows chemists and DMPK scientists to optimize molecules across multiple dimensions, including systemic exposure!
And now, starting in APX3, communication with 3rd-party software is also available. This unlocks incredible flexibility with the engine -utilize your 3D docking scores from tools like GLIDE, GOLD, AutoDock, or predictions from other tools, with our machine learning/PBPK models and risk liability scores as part of the optimization.
The AIDD Module integrates ADMET Predictor’s
The AIDD Module integrates ADMET Predictor’s top-ranked ADMET property prediction models and high throughput pharmacokinetic (HTPK) simulations with multi-objective compound optimization capabilities:
- Generative chemistry engine
- Advanced and customizable pruning rules
- Multi-objective compound selection based on a Pareto front
- Iterative cycles of optimization based on design criteria
Compound generation and evaluation against target profile
Taking advantage of these capabilities, AIDD can generate and evaluate up to 10 million molecules during an overnight run (8-core laptop computer), exploring a large portion of chemical space around the compound(s) or scaffold(s) of interest.
AIDD also empowers chemistry and DMPK teams to control which part(s) of the molecule may be altered as part of the optimization and which should be maintained. You can also specify positions where substitutions can be applied, and control the chemistry based on synthetic feasibility constraints or a-priori knowledge about the target.
A full suite of cheminformatics functionality is also included to provide you with the capabilities to visualize and analyze results from AIDD runs at any stage of the design process.
Best of all?
The AIDD Module provides a holistic user experience built on top of the easy-to-use ADMET Predictor software. Nothing new to learn – only intuitive interpretation and interaction with the results is necessary.
#1-ranked machine learning ADMET property + #1-ranked PBPK models, wrapped in proven generative chemistry approaches… only in the AIDD Module.