Related Experiment Video
Updated: May 9, 2026
10:12
Shotgun Proteomics Sample Processing Automated by an Open-Source Lab Robot
Published on: October 28, 2021
ScopeMap: An AI-Assisted, Human-in-the-Loop Workflow for Mapping Reaction Scope and Boundaries
Jiawei Li1,2, Xiao Xiao3, Qi Yang1,2,4
1Center of Basic Molecular Science (CBMS), Department of Chemistry, Tsinghua University, Beijing, China.
Angewandte Chemie (International Ed. in English)
|May 8, 2026
Summary
ScopeMap efficiently maps synthetic reaction limits using a human-in-the-loop approach. This method explores reaction boundaries more effectively than traditional algorithms, improving the assessment of chemical synthesis generality.
Area of Science:
- Organic Chemistry
- Computational Chemistry
- Chemical Synthesis
Background:
- Assessing the scope of synthetic methods is crucial in organic chemistry.
- Current methods often focus on optimization, failing to define reaction boundaries.
- This leads to clustering in high-reactivity areas, limiting understanding of generality.
Purpose of the Study:
- To introduce ScopeMap, an iterative workflow for mapping functional limits of synthetic reactions.
- To move beyond performance maximization towards efficient boundary delineation.
- To provide a resource-efficient paradigm for defining reaction generality.
Main Methods:
- Developed ScopeMap, a human-in-the-loop workflow utilizing a modified Centroidal Voronoi Tessellation (CVT) algorithm.
- Employed a dynamic geometric repulsion potential to steer sampling toward unexplored frontiers.
- Introduced U-Score and R-Score metrics for quantifying sampling uniformity and representativeness.
Main Results:
- ScopeMap achieved greater substrate diversity with fewer examples in aldol reactions and cobalt-catalyzed coupling.
- A representative subset (<4% of substrate space) predicted reactivity with >90% F1 score.
- The workflow successfully transformed negative feedback into geometric constraints for boundary mapping.
Conclusions:
- ScopeMap offers an efficient method for mapping the functional limits of chemical reactions.
- This approach shifts focus from exhaustive data to information-dense boundary mapping.
- The U-Score and R-Score provide a standardized framework for evaluating sampling strategies.
