Molecule Set Generation via Pre-screening and Biochemical Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The de novo design of drugs faces challenges due to the vast chemical space of drug-like molecules, making it difficult to efficiently generate molecules with desirable pharmacological properties, as existing methods are costly, time-consuming, and lack effective evaluation processes.
Innovation Solution
A method and electronic device for generating a molecule set using a pre-screening model to obtain an initialization molecule subset, acquire physical information, screen molecules based on this information, and obtain a target molecule set through biochemical experimental evaluation, incorporating techniques like genetic algorithms and neural networks to improve efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional exhaustive search methods are used to find molecules with desirable properties in the vast chemical space, then comprehensive coverage of chemical space is achieved, but the computational cost and time consumption become prohibitively high
Solution Approach 1:
The patent segments the vast chemical space into manageable subsets by generating molecules through controlled modification of seed molecules. Instead of searching the entire chemical space (10^60-10^100 possibilities), the system divides it into regions accessible through systematic molecular transformations, making the search tractable while maintaining comprehensive coverage of relevant chemical space.
Solution Approach 2:
The patent applies preliminary action by using pre-screening models to evaluate and filter molecules before full biochemical experimental evaluation. The pre-screening model performs initial assessments of molecular properties, eliminating clearly inferior candidates early in the process, thereby reducing the number of molecules that require time-consuming experimental verification.
2Measurement precision
If comprehensive biochemical experimental evaluation is performed on all generated molecules, then accurate assessment of pharmacological properties is achieved, but the cost and time requirements become excessive
Solution Approach 1:
The patent implements preliminary action through a two-stage evaluation process. First, a pre-screening model performs rapid computational assessment of molecular properties including binding affinity, solubility, and toxicity. Only molecules that pass this preliminary filter proceed to full biochemical experimental evaluation, significantly reducing the number of experiments needed while maintaining accuracy for promising candidates.
Solution Approach 2:
The pre-screening model serves as an intermediary between molecular generation and full biochemical evaluation. This intermediary layer filters and prioritizes molecules based on computational predictions, allowing the system to focus experimental resources on the most promising candidates while maintaining measurement precision for the final selection.
3Reliability
If diverse molecular structures are generated to explore different chemical spaces, then the probability of finding optimal molecules increases, but the complexity of the generation process and computational resources required increase
Solution Approach 1:
The patent applies universality through a multi-functional molecular generation system that can perform multiple operations using the same core infrastructure. The system can generate diverse molecules through various modification strategies (addition, deletion, substitution of molecular groups) while using a unified computational framework, reducing overall process complexity compared to maintaining separate generation systems for each molecular type.
Solution Approach 2:
The patent utilizes parameter changes to generate molecular diversity by systematically varying molecular properties such as the type and position of functional groups, molecular size, and structural complexity. By controlling these parameters through defined transformation rules, the system explores diverse chemical space while maintaining manageable process complexity through systematic parameter management.
Data Source
AI summary
Embodiments of the present disclosure provide a method and electronic device for generating a molecule set and a storage medium thereof. The method obtains the first initialization molecule subset from the initialization molecule set with the pre-screening model; acquires the physical information of at least one initialization molecule in the first initialization molecule subset, and screens at least one initialization molecule based on the physical information to obtain the screened molecule set; acquires the biochemical experimental evaluation value of at least one molecule in the screened molecule set; and obtains the target molecule set based on the biochemical experimental evaluation value of at least one molecule.