Cycle-Configuration: A Novel Graph-theoretic Descriptor Set for Molecular Inference

TL;DR

This paper introduces cycle-configuration descriptors for chemical graphs, enhancing molecular inference by capturing aromatic ring patterns, leading to improved property prediction and feasible graph inference within the 2L model.

Contribution

It proposes a new family of graph descriptors called cycle-configuration, enabling better molecular property prediction and practical inference in the 2L framework.

Findings

Descriptors improve prediction accuracy for 27 chemical properties.

The MILP formulation can infer graphs with up to 50 vertices efficiently.

Cycle-configuration captures aromatic ring patterns previously unrepresented.

Abstract

In this paper, we propose a novel family of descriptors of chemical graphs, named cycle-configuration (CC), that can be used in the standard "two-layered (2L) model" of mol-infer, a molecular inference framework based on mixed integer linear programming (MILP) and machine learning (ML). Proposed descriptors capture the notion of ortho/meta/para patterns that appear in aromatic rings, which has been impossible in the framework so far. Computational experiments show that, when the new descriptors are supplied, we can construct prediction functions of similar or better performance for all of the 27 tested chemical properties. We also provide an MILP formulation that asks for a chemical graph with desired properties under the 2L model with CC descriptors (2L+CC model). We show that a chemical graph with up to 50 non-hydrogen vertices can be inferred in a practical time.