Differentiable Molecule Generation and Docking for Multi-Property Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computational chemistry models are limited in designing molecules that satisfy multiple desired physicochemical properties, particularly in the context of computer-aided drug discovery, due to the vastness of the chemical space and high computational cost.
Innovation Solution
A method combining generative modeling with multi-objective optimization, using differentiable machine learning models to guide the generation of ligands while optimizing for multiple target properties such as binding affinity and synthetic accessibility, and employing a series of machine learning models for structure docking, scoring, and updating candidate structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If traditional database screening methods are used to evaluate molecules, then the process is straightforward and computationally manageable, but the chemical space explored remains extremely limited (billions out of 10^20 to 10^60 possible molecules)
Solution Approach 1:
The patent replaces traditional mechanical database screening methods with a generative AI model that can explore chemical space through learned patterns and transformations. The model generates novel molecular structures by applying geometric transformations to input molecules, enabling exploration of previously inaccessible chemical space without proportional increases in computational cost
Solution Approach 2:
The patent changes the parameter space by working in 3D geometric space with SE(3) equivariant transformations rather than traditional 2D molecular representations. This allows the model to generate structurally diverse molecules with controlled 3D conformations, expanding the effective chemical space explored while maintaining computational efficiency
2Reliability
If the search chemical space is expanded to discover better materials, then the likelihood of finding optimal molecules increases, but the computational cost and time requirements increase dramatically
Solution Approach 1:
The patent performs preliminary action by pre-training the generative model on existing molecular databases to learn fundamental chemical patterns and relationships. This pre-learning enables the model to generate high-quality novel molecules with fewer computational resources during the actual drug discovery process, as the heavy lifting of pattern recognition has already been performed
Solution Approach 2:
The patent uses copying by generating novel molecules through transformations of existing molecular structures rather than performing exhaustive searches. The model learns from copies of known molecules and generates new structures by applying learned transformations, achieving efficient exploration of chemical space without proportional computational cost increases
Data Source
AI summary
A method for machine learning aided modeling of two interacting structures may include: (a) receiving an input structure comprising an interaction region; (b) generating a plurality of candidate structures using a first differentiable machine learning model; (c) docking one or more candidate structures of the plurality of candidate structures at the interaction region of the input structure using a second differentiable machine learning model to predict a docking geometry; (d) ranking the one or more candidate structures of the plurality of candidate structures docked in (c) using a third differentiable machine learning model to predict a score; and (e) backpropagating the score to (i) the first differentiable machine learning model to update the plurality of candidate structures or (ii) the second differentiable machine learning model to update the docking geometry.