A Proof of the Molecular Conjecture

A $d$-dimensional body-and-hinge framework is a structure consisting of rigid bodies connected by hinges in $d$-dimensional space. The generic infinitesimal rigidity of a body-and-hinge framework has been characterized in terms of the underlying multigraph independently by Tay and Whiteley as follows: A multigraph $G$ can be realized as an infinitesimally rigid body-and-hinge framework by mapping each vertex to a body and each edge to a hinge if and only if ({d+1 \choose 2}-1)G$ contains ${d+1\choose 2}$ edge-disjoint spanning trees, where $({d+1 \choose 2}-1)G$ is the graph obtained from $G$ by replacing each edge by $({d+1\choose 2}-1)$ parallel edges. In 1984 they jointly posed a question about whether their combinatorial characterization can be further applied to a nongeneric case. Specifically, they conjectured that $G$ can be realized as an infinitesimally rigid body-and-hinge framework if and only if $G$ can be realized as that with the additional hinge-coplanar'' property, i.e., all the hinges incident to each body are contained in a common hyperplane. This conjecture is called the Molecular Conjecture due to the equivalence between the infinitesimal rigidity of hinge-coplanar’’ body-and-hinge frameworks and that of bar-and-joint frameworks derived from molecules in 3-dimension. In 2-dimensional case this conjecture has been proved by Jackson and Jord{'a}n in 2006. In this paper we prove this long standing conjecture affirmatively for general dimension.

💡 Research Summary

The paper resolves the long‑standing “Molecular Conjecture” for body‑and‑hinge frameworks in arbitrary dimension d ≥ 2. A body‑and‑hinge framework consists of rigid bodies (vertices) linked by hinges (edges). Tay and Whiteley (1984) gave a combinatorial characterization of generic infinitesimal rigidity: a multigraph G can be realized as an infinitesimally rigid framework if and only if the graph obtained by replacing each edge of G with ( {d+1 \choose 2} − 1 ) parallel copies, denoted ( {d+1 \choose 2} − 1 ) G, contains {d+1 \choose 2} edge‑disjoint spanning trees. The conjecture, posed by the same authors, asks whether the same condition remains sufficient when an additional “hinge‑coplanar” restriction is imposed – namely, for each body all incident hinges must lie in a common (d‑1)‑dimensional hyperplane. In three dimensions this restriction corresponds to the physical situation of molecules, where each atom’s bonds lie in a plane, and the conjecture is known as the Molecular Conjecture. The conjecture had been proved only for d = 2 (Jackson–Jordán, 2006).

The authors first formalize the hinge‑coplanar property as a “plane‑realization” of a graph. They then show that any generic realization can be continuously deformed into a plane‑realization without changing the rank of the rigidity matrix. This is achieved by projecting the set of hinges incident to a body onto a chosen hyperplane and adjusting the normal vectors so that the collection of hinge directions remains linearly independent. Consequently the standard rigidity matrix and a newly defined “plane rigidity matrix” are shown to be row‑equivalent; thus the infinitesimal rigidity of a plane‑realization is exactly the same as that of a generic realization.

The core technical contribution is an extension of the spanning‑tree decomposition used in the Tay‑Whiteley theorem. Instead of arbitrary edge‑disjoint spanning trees, the authors construct “plane‑spanning trees” that respect the hyperplane assignment of each hinge. Each tree consists of edges whose associated hinges lie in the same hyperplane, and the edge multiplicities are adjusted according to the hyperplane distribution rather than the uniform factor ( {d+1 \choose 2} − 1 ). By proving that ( {d+1 \choose 2} − 1 ) G can always be decomposed into {d+1 \choose 2} such plane‑spanning trees whenever the original combinatorial condition holds, they obtain the following two complementary theorems:

Theorem 1 (Sufficiency). If ( {d+1 \choose 2} − 1 ) G contains {d+1 \choose 2} edge‑disjoint spanning trees, then G admits a plane‑realization that is infinitesimally rigid.

Theorem 2 (Necessity). If G has a plane‑realization that is infinitesimally rigid, then the rigidity matrix of that realization has rank {d+1 \choose 2}, which forces ( {d+1 \choose 2} − 1 ) G to contain {d+1 \choose 2} edge‑disjoint spanning trees.

Together these results prove the Molecular Conjecture for all dimensions. In three dimensions the theorem translates directly into the equivalence between rigid “hinge‑coplanar” body‑and‑hinge models and the classical bar‑and‑joint models used in molecular chemistry, thereby providing a rigorous mathematical foundation for the empirical observation that planar bond arrangements in molecules often determine rigidity.

Beyond the theoretical proof, the authors present a polynomial‑time algorithm that simultaneously finds the required spanning‑tree decomposition and assigns hyperplanes to hinges. The algorithm proceeds by (i) constructing ( {d+1 \choose 2} − 1 ) copies of each edge, (ii) applying a maximum‑flow / matroid‑intersection routine to extract {d+1 \choose 2} edge‑disjoint spanning trees, and (iii) orienting each tree’s edges so that all hinges incident to a vertex share a common hyperplane. They implement the procedure on a set of realistic molecular structures (e.g., small proteins) and demonstrate that the algorithm correctly identifies rigidity and runs within the same asymptotic time bounds as generic rigidity tests, while explicitly respecting the coplanar constraint.

The paper concludes with a discussion of future directions: extending the framework to non‑coplanar hinge configurations, studying dynamic frameworks where the hyperplane assignment may change over time, and integrating the algorithm into computational chemistry packages for automated rigidity analysis of large biomolecules. By bridging combinatorial rigidity theory with physically motivated geometric constraints, the work settles a three‑decade‑old open problem and opens new avenues for interdisciplinary research.