Fact-checked by Grok 2 weeks ago

Molecular graph

A molecular graph, also referred to as a chemical graph, is an abstract representation of a molecule's structural formula in graph theory, where vertices correspond to atoms and edges represent covalent chemical bonds between them.^[1] This model captures the topological connectivity of the molecule, typically for constitutional (2D) structures, while abstracting away stereochemical or three-dimensional spatial details unless explicitly incorporated.^[2] Nodes in the graph may be labeled or weighted by atomic properties such as element type, and edges by bond orders (e.g., single, double, or aromatic).^[3] The foundations of molecular graphs and chemical graph theory emerged in the 19th century, building on early atomic models by John Dalton in 1808^[4] and graphical bond depictions by Archibald Couper in 1858.^[5] Arthur Cayley advanced the field in 1874 by enumerating tree-like structures for alkane hydrocarbons, laying groundwork for isomer counting, while James Sylvester coined the term "graph" in 1878 to describe molecular connectivity.^[5]^[6] Key formalization occurred in the 20th century with George Pólya's 1937 enumeration theorem, which used group theory to systematically generate and count molecular isomers, such as those of alkanes or annulenes.^[2] The discipline gained momentum post-1970s through computational tools and the introduction of topological indices, enabling quantitative links between graph structure and molecular properties.^[7] Molecular graphs underpin numerous applications in cheminformatics and computational chemistry, including the enumeration and coding of constitutional, steric, and valence isomers for compound generation and validation.^[2] They facilitate database searching and similarity assessments via graph invariants, such as the Wiener index (sum of shortest path distances between all atom pairs, correlating with molecular branching and boiling points) and the Randić connectivity index (edge-weighted by inverse square root of vertex degrees, used in quantitative structure-activity relationships or QSAR).^[7] In organic synthesis, synthon and reaction graphs model multistep pathways, aiding computer-assisted planning tools like CONGEN or modern successors such as MAYGEN (2021).^[5] Additionally, molecular graphs support computer-assisted structure elucidation (CASE) in metabolomics, where fragmentation algorithms (e.g., MetFrag) reconstruct unknown compounds from mass spectrometry data, and enable machine learning models for drug discovery and property prediction.^[5]

Fundamentals

Definition

A molecular graph is a representation of a chemical compound's structure as an undirected graph, where vertices correspond to atoms and edges correspond to chemical bonds between them.^[8]^[9] This model abstracts the molecule's topology, focusing on connectivity rather than physical attributes.^[10] Unlike more complex graph models that may include directions or multiple edges, molecular graphs are typically simple and undirected, reflecting the bidirectional nature of covalent bonds without self-loops or parallel edges.^[9] They are unweighted in their basic form, though edge weights or labels can be added to denote bond orders (e.g., single, double, or triple bonds) when needed for specific analyses.^[10] The vertices may carry labels to indicate atomic types, such as carbon or hydrogen, but the core structure emphasizes adjacency over quantitative bond strengths.^[8] Fundamental assumptions in this representation treat atoms as nodes devoid of inherent spatial geometry, and bonds as mere connections that disregard stereochemical configurations like chirality or cis-trans isomerism.^[9] This topological approach simplifies molecular analysis by prioritizing combinatorial properties over three-dimensional arrangements. For instance, the molecular graph of methane (CH₄) forms a star graph, or K_{1,4}, with the central carbon vertex connected to four peripheral hydrogen vertices, illustrating a basic tree-like structure without cycles.^[11]

Representation

Molecular graphs are visually represented by adapting Lewis structures into graph notation, where atoms serve as vertices and chemical bonds as edges, often omitting explicit hydrogen atoms for carbon-centered skeletons to emphasize connectivity. This depiction simplifies complex molecular structures into skeletal formulas, facilitating qualitative analysis of topology in chemical graph theory. For small molecules, adjacency matrices provide a quantitative visual and computational representation, with rows and columns corresponding to atoms and entries indicating bond presence (typically 1 for single bonds between connected atoms, 0 otherwise).^[12] In computational encodings, SMILES (Simplified Molecular Input Line Entry System) strings linearize molecular graphs into text-based notations that capture atom types, connectivity, and stereochemistry, interpretable as graphs where sequential symbols denote vertices and implied or explicit symbols represent edges.^[13] Canonical labeling ensures unique representations by assigning consistent atom orders based on graph isomorphism algorithms, enabling standardized comparisons and storage of molecular structures across databases.^[13] This process involves generating multiple possible labelings and selecting the lexicographically smallest one, often integrated into SMILES generation for reproducibility.^[14] Multiple bonds, such as double or triple bonds, are handled in molecular graphs through multigraphs, where parallel edges between vertices represent bond multiplicity, or via edge labels assigning weights (e.g., 2 for double bonds).^[15] This extension from simple graphs preserves valence information essential for reactivity predictions. For example, benzene is represented as a cycle graph C_6 with six carbon vertices forming a ring and alternating double bonds encoded as edge weights of 1 and 2, or in SMILES as c1ccccc1 where lowercase 'c' implies aromaticity and single/double bonds are delocalized.^[13]

Structural Properties

Valence and Degree

In chemical graph theory, molecular graphs are typically represented as hydrogen-depleted structures, where vertices correspond to non-hydrogen atoms and edges represent covalent bonds between them, excluding bonds to hydrogen atoms. The degree of a vertex in such a graph is defined as the number of edges incident to it, which quantifies the bonds formed by the atom with other non-hydrogen atoms. This degree directly relates to the chemical valence of the atom but accounts only for explicit connections in the graph, necessitating adjustments for implicit hydrogens to fully satisfy valence rules.^[16]^[17] The valence of an atom denotes its capacity to form bonds, determined by its electron configuration in organic molecules. For instance, carbon atoms have a standard valence of 4, oxygen atoms 2, and nitrogen atoms 3, reflecting their typical bonding patterns in neutral organic compounds. In the hydrogen-depleted molecular graph, the vertex degree for a heteroatom like oxygen is at most 2 (e.g., in ethers or alcohols), while for nitrogen it is at most 3 (e.g., in amines), ensuring compatibility with these valence constraints. Molecular graphs may use edge labels or weights to represent bond orders (single, double, triple, or aromatic). Implicit hydrogens are calculated for each vertex as the difference between the atom's standard valence and the sum of the bond orders of its connections to other non-hydrogen atoms, providing a complete bonding profile without explicitly including hydrogen vertices. This approach simplifies graph-based computations while preserving chemical accuracy.^[18]^[19] A representative example is the molecular graph of ethanol (C₂H₅OH), where the vertices are two carbon atoms (C1 for the methyl group, C2 for the methylene group) and one oxygen atom, with edges connecting C1 to C2 and C2 to O. The degree of C1 is 1 (bonded only to C2), so it requires 3 implicit hydrogens to reach a valence of 4; C2 has degree 2 (bonded to C1 and O), requiring 2 implicit hydrogens; and O has degree 1 (bonded to C2), requiring 1 implicit hydrogen to satisfy its valence of 2. This structure yields the full formula C₂H₆O, illustrating how graph degrees and valence rules reconstruct the molecule's hydrogen content.^[20]^[12]

Connectivity and Cycles

In molecular graphs, connectivity refers to the existence of paths between vertices representing atoms, where a path is a sequence of distinct edges (bonds) connecting two vertices without repetition. A molecular graph is connected if there is at least one path between every pair of vertices, which corresponds to a single, intact molecule where all atoms are linked through covalent bonds. In contrast, disconnected molecular graphs represent separate molecular components, such as in mixtures or dissociated species, but single molecules are typically modeled as connected graphs to reflect their structural unity.^[21] Cycles in molecular graphs capture ring structures, which are closed paths where the starting and ending vertices coincide, forming loops of atoms and bonds essential for molecular stability and reactivity. For instance, cyclohexane is represented as a simple 6-cycle (C6) in its molecular graph, with six carbon vertices each connected to two neighbors, embodying the saturated ring's topology. These cycles distinguish cyclic compounds from acyclic ones, influencing properties like planarity and strain.^[21] Bridge edges, or cut edges, are bonds whose removal increases the number of connected components in the molecular graph, indicating critical linkages that, if broken, would fragment the molecule into separate parts. Similarly, articulation points (cut vertices) are atoms whose removal, along with their incident edges, disconnects the graph, highlighting pivotal sites for structural integrity, such as branch points in complex hydrocarbons. In naphthalene, a fused cycle graph consists of two 6-cycles sharing two adjacent vertices and the edge between them, forming a bicyclic structure without bridge edges or articulation points, ensuring high connectivity. Vertex degrees, as endpoints in paths, further contextualize these features by quantifying local bonding at cycle junctions.^[22]^[21]

Topological Invariants

Graph Invariants

Graph invariants, also known as topological indices in chemical graph theory, are numerical quantities derived from the structure of a molecular graph that remain unchanged under graph isomorphisms and provide summaries of molecular topology. These invariants capture essential features such as size, connectivity, and branching without reference to atomic coordinates or three-dimensional geometry, making them valuable for comparative analysis in chemistry. In molecular graphs, where vertices represent atoms and edges represent bonds, invariants are typically computed on hydrogen-suppressed graphs to focus on the carbon skeleton or full atomic structure as needed. The most fundamental graph invariants are the number of vertices, edges, and connected components. The number of vertices, denoted |V(G)|, equals the number of atoms in the molecule, while the number of edges |E(G)| corresponds to the number of bonds; for example, in a saturated hydrocarbon graph, |E(G)| = (|V(G)| - 1) due to the tree-like structure.^[2] The number of connected components indicates whether the molecular graph represents a single molecule or multiple entities, such as in salts or reaction mixtures, with most neutral organic molecules having one component. Distance-based invariants quantify the overall spread or compactness of the molecular structure using shortest path distances between atoms. The Wiener index, introduced by Harry Wiener in 1947, is defined as the sum of the shortest path distances over all unordered pairs of vertices:

W(G) = \sum_{1 \leq i < j \leq n} d(i,j)

where d(i,j) is the length of the shortest path between vertices i and j, and n = |V(G)|. This index measures the "size" of the molecule in topological terms and correlates with properties like boiling points in alkanes, as Wiener originally demonstrated. Branching and shape invariants account for the degree distribution and skeletal complexity of molecules. The Randić index, proposed by Milan Randić in 1975, is a connectivity index that weights edges by the inverse square root of the product of the degrees of their incident vertices:

\chi(G) = \sum_{(i,j) \in E(G)} (d_i d_j)^{-1/2}

where d_i and d_j are the degrees (valences) of vertices i and j. This invariant penalizes highly branched structures with low-degree vertices and has been widely used to distinguish isomers with similar vertex counts but different topologies. For illustration, consider n-pentane (C_5H_{12}), whose molecular graph is a path of five vertices. The pairwise distances are: four at distance 1, three at 2, two at 3, and one at 4, yielding W(G) = 1 \cdot 4 + 2 \cdot 3 + 3 \cdot 2 + 4 \cdot 1 = 4 + 6 + 6 + 4 = 20. This value is higher than for branched pentane isomers like neopentane (W=16), highlighting how linear chains extend topological distances.

Molecular Descriptors

Molecular descriptors in the context of molecular graphs are numerical invariants derived from graph structures to predict physicochemical properties such as reactivity, stability, and energy levels of molecules. These descriptors extend basic graph invariants by incorporating chemical-specific features like valence and bonding patterns, enabling quantitative structure-property relationships (QSPR) in computational chemistry. Unlike purely topological measures, they often integrate spectral or degree-based information to model molecular behavior more accurately.^[23] Topological indices for reactivity, such as the Hosoya index, quantify the number of independent edge matchings in a molecular graph, providing insights into molecular branching and saturation effects on reactivity. Introduced by Haruo Hosoya, the index Z(G) for a graph G is defined as the total number of matchings, given by the evaluation of the matching polynomial at x=1, or equivalently Z(G) = \sum_{k=0}^{\lfloor n/2 \rfloor} m_k(G), where m_k(G) is the number of k-edge matchings and n is the number of vertices. For trees, it satisfies the recursion Z(G) = Z(G - v) + Z(G - u - v), where v is a pendant vertex adjacent to u. This index correlates with properties like the heat of formation and reactivity in saturated hydrocarbons, where higher Z values indicate greater stability due to more dispersed matchings. Spectral descriptors leverage the eigenvalues of the adjacency matrix A of a molecular graph to capture vibrational and electronic properties, approximating molecular orbitals in quantum chemistry models. The characteristic polynomial is defined as P(G, \lambda) = \det(\lambda I - A), whose roots are the eigenvalues \lambda_1 \geq \lambda_2 \geq \cdots \geq \lambda_n, providing a spectrum that encodes symmetry and connectivity. These eigenvalues relate to the Hückel molecular orbital energies in conjugated systems, where the largest eigenvalue \lambda_1 approximates the highest occupied molecular orbital energy, influencing reactivity. Spectral moments, such as \sum \lambda_i^k, serve as descriptors for QSAR, correlating with properties like polarizability in organic molecules.^[23]^[24] Valence-specific descriptors, like the Zagreb indices, emphasize vertex degrees to model bonding strength and molecular compactness, particularly in hydrocarbons where degree represents valence. The first Zagreb index is M_1(G) = \sum_{i=1}^n d_i^2, summing the squares of degrees d_i over all vertices, while the second is M_2(G) = \sum_{uv \in E(G)} d_u d_v, summing products of degrees over edges. Developed by Ivan Gutman and Nenad Trinajstić, these indices approximate total π-electron energy in alternant hydrocarbons, with M_1 capturing overall branching and M_2 edge-wise interactions.^[25]

Applications

Chemical Modeling

Molecular graphs serve as abstract representations of chemical structures, enabling simulations and analyses that abstract away from explicit atomic coordinates while preserving connectivity information. In chemical modeling, these graphs facilitate the exploration of molecular behaviors through combinatorial and algorithmic approaches, allowing chemists to predict structural motifs, spatial arrangements, and reactivity patterns without relying on computationally intensive geometric optimizations. This graph-centric paradigm is particularly valuable for high-throughput screening and hypothesis generation in drug design and materials science.^[26] Graph-based structure search employs subgraph isomorphism algorithms to identify specific functional groups within larger molecular graphs, treating the target motif—such as a hydroxyl or carbonyl—as a query subgraph to match against the host molecule. This method efficiently detects embedded patterns by mapping vertices and edges while respecting labels for atom types and bond orders, reducing the complexity of manual inspection in complex scaffolds. For instance, Ullmann's algorithm and its variants have been benchmarked for substructure matching in molecular databases, demonstrating high accuracy for common functional groups like aromatic rings.^[26] Topological indices derived from these graphs can further quantify similarity, aiding in prioritization of matches.^[27] In conformational analysis, molecular graphs act as topological scaffolds that guide the embedding of atoms into three-dimensional space, decoupling connectivity from initial geometric constraints to explore viable low-energy conformers. By representing the molecule's bonding network, graphs enable systematic enumeration of possible spatial arrangements, such as ring puckering or chain folding, through distance geometry or generative models that assign coordinates while minimizing steric clashes. Conditional generative neural networks, for example, use graph inputs to produce diverse 3D conformations tailored to specified properties like chirality or flexibility, bypassing exhaustive sampling.^[28] This approach is essential for modeling flexible biomolecules where multiple conformations influence binding affinity. Reaction mapping leverages graph editing operations to simulate chemical transformations, where bond breaking and forming are depicted as edge deletions, additions, or label changes, thereby tracing reaction pathways from reactants to products. Algorithms based on graph edit distance compute minimal sequences of such operations, capturing the stoichiometry and connectivity shifts in organic syntheses. This framework allows for the prediction of feasible routes by minimizing edit costs, as demonstrated in solid-state reaction networks.^[29]

Computational Chemistry

In computational chemistry, molecular graphs serve as foundational structures for various algorithms in cheminformatics, enabling efficient analysis of molecular connectivity and spatial relationships. Shortest path algorithms, such as Floyd-Warshall or Dijkstra's, are commonly employed to compute distance metrics between atoms, which underpin topological descriptors like the Wiener index that quantify overall molecular branching and compactness. These metrics facilitate tasks such as similarity searching and pharmacophore identification by measuring the minimum number of bonds between atoms. For tree-like structures such as dendrimers, which are modeled as acyclic graphs with a central core and branching layers, tree traversal algorithms like depth-first search are utilized to enumerate branches, calculate generation-specific distances, and derive structural invariants for property prediction.^[9]^[30]^[31]^[32]^[33] Quantitative structure-activity relationship (QSAR) and quantitative structure-property relationship (QSPR) models leverage molecular graph descriptors to perform regression analyses for predicting physicochemical properties, such as normal boiling points, which correlate with intermolecular forces and molecular size. In these models, graph-based features like eccentricity or path counts are fed into linear or nonlinear regressors to estimate boiling points for diverse organic compounds. Topological indices have been shown to provide good predictive accuracy for boiling points of alkanes and other compounds.^[34]^[35]^[36] Key software tools for molecular graph manipulation include RDKit and OpenBabel, both open-source cheminformatics libraries that represent molecules as editable graphs for tasks like substructure matching and descriptor computation. RDKit, implemented in C++ with Python bindings, supports graph-based operations such as adjacency matrix generation and bond perception, while OpenBabel excels in format interconversion and graph canonicalization for large datasets. These tools integrate seamlessly with molecular dynamics (MD) simulations; for example, RDKit can generate 3D conformations and force field parameters compatible with OpenMM, enabling hybrid workflows where graph-derived topologies initialize MD trajectories for energy minimization and property sampling.^[37]^[38]^[39]^[40]^[41] A prominent example of advanced graph-based prediction involves graph neural networks (GNNs), which process molecular adjacency matrices to forecast properties like energy levels or reactivity. In the seminal work on neural message passing, GNNs propagate features along graph edges to learn representations that outperform traditional descriptors in quantum chemistry tasks, such as predicting dipole moments with mean absolute errors around 0.2 Debye for small molecules. This approach treats the adjacency matrix as input to iterative message updates, capturing long-range interactions without explicit coordinate geometry. Recent advances as of 2025 have further integrated graph representations with generative models for de novo molecule design in drug discovery.^[42]^[43]

Historical Development

Origins in Graph Theory

The origins of molecular graph theory lie in the foundational developments of graph theory, which provided the mathematical framework for modeling connectivity and structure in abstract networks. The field of graph theory itself emerged from Leonhard Euler's seminal 1736 solution to the Seven Bridges of Königsberg problem, where he demonstrated that no continuous path could traverse all seven bridges exactly once due to the odd degrees of vertices in the resulting multigraph. This work introduced key concepts of connectivity and Eulerian paths, establishing graph theory as a tool for analyzing path traversability in networks, which later proved essential for representing molecular bonds as edges between atomic vertices. Euler's approach abstracted geographical features into vertices and edges, laying the groundwork for applying similar representations to discrete structures like molecules.^[44] A significant advancement came in 1875 with Arthur Cayley's enumeration of tree structures, directly linking graph theory to chemical isomer counting. Cayley formalized the enumeration of unlabeled trees with up to four branches per vertex, corresponding to the carbon atoms in alkane molecules (CnH2n+2), where trees represent acyclic hydrocarbon skeletons without cycles. His method used generating functions to count distinct isomers, such as the 18 possible structures for C8H18, by considering rooted and unrooted trees while accounting for valence constraints. This application marked one of the earliest uses of graph-theoretic enumeration in chemistry, treating molecular formulas as constraints on tree valency.^[45] Building on such enumerative techniques, George Pólya developed his enumeration theorem in 1937, incorporating group theory to count distinct graphs under symmetry operations. Pólya's theorem employs the cycle index of a permutation group acting on graph colorings or labelings to generate the number of non-isomorphic structures, applicable to counting molecular graphs invariant under rotations or reflections. For instance, it efficiently enumerates benzene derivatives by averaging fixed points over the dihedral group symmetries, avoiding redundant listings of equivalent configurations. This group-theoretic approach revolutionized combinatorial graph counting, providing a systematic way to handle symmetries inherent in molecular representations.^[46] Initial explicit applications of these graph-theoretic ideas to chemical structures appeared in Frank Harary and Robert Z. Norman's 1953 work on Husimi trees, which extended tree dissimilarity characteristics to clustered graphs modeling molecular assemblies. Husimi trees, defined as graphs where every edge lies on exactly one cycle, captured branching patterns in polyatomic molecules, with dissimilarity measures quantifying structural differences between isomers. Their analysis introduced graphical invariants like the dissimilarity characteristic polynomial to distinguish non-isomorphic chemical graphs, bridging pure graph enumeration with practical molecular identification. This paper formalized the use of graph theory for chemical modeling, influencing subsequent developments in topological descriptors.^[47]

Evolution in Chemistry

The application of graph theory to molecular structures gained traction in chemistry during the 1960s, particularly through the work of Alexandru T. Balaban, who pioneered the use of graphs to enumerate isomers of cage hydrocarbons and annulenes. Balaban's 1966 publication introduced reaction graphs to model elementary steps in organic synthesis, marking an early chemical adaptation of graph-theoretic tools for structural analysis beyond pure mathematics. This laid groundwork for topological indices tailored to chemical compounds, with Balaban later developing the Balaban index in 1982 specifically for distinguishing cage-like structures in quantitative structure-activity relationship (QSAR) studies.^[48]^[49] In the 1970s and 1980s, chemists revived and expanded early topological invariants like the Wiener index—originally proposed in 1947 for predicting alkane properties—for QSAR modeling, correlating molecular branching with biological activities such as anesthesia potency. This revival emphasized distance-based descriptors to quantify structural complexity, enabling predictive regressions for physicochemical properties. A seminal contribution was Milan Randić's 1975 connectivity index, which weighted bonds by vertex degrees to better capture branching effects, achieving superior correlations in QSAR for diverse organic compounds compared to unweighted paths. The index's influence persisted, underpinning thousands of subsequent QSAR applications by the 1980s.^[50]^[51] A key milestone in the late 1980s was the development of the Simplified Molecular Input Line Entry System (SMILES) by David Weininger in 1988, a string-based notation for serializing molecular graphs that facilitated machine-readable representation of chemical structures. SMILES encoded atoms as symbols and bonds as connections, enabling efficient storage and querying in computational workflows. From the 1990s onward, molecular graphs integrated with bioinformatics, modeling protein interaction networks and metabolic pathways as graphs to analyze genomic data from projects like the Human Genome Project. This era saw graph databases emerge for large-scale chemical screening, allowing substructure searches across millions of compounds to identify drug candidates, with systems like those based on Neo4j later optimizing virtual screening pipelines.^[13]^[52]^[53]^[54]

References

[1]
Molecular graph – Knowledge and References - Taylor & Francis
A molecular graph is a visual representation of a molecule, where the vertices represent the atoms and the edges represent the chemical bonds between them.
[2]
[PDF] Applications of Graph Theory in Chemistry
Con- stitutional (molecular) graphs have points (vertices) representing atoms and lines (edges) sym- bolizing malent bonds. This review deals with definition.
[3]
Molecule Graph - Chemaxon Docs
A molecule graph represents chemical structures where atoms are nodes and bonds are edges. Nodes can be colored with atom types and edges with bond types.
[4]
[PDF] Review on Chemical Graph Theory and Its Application in Computer ...
Nov 29, 2021 · This review first covers the history of chemical graph theory, then provides an overview of its various techniques and applications for CASE, ...
[5]
Applications of graph theory in chemistry - ACS Publications
Application of Chemical Graph Theory for Automated Mechanism Generation. Journal of Chemical Information and Computer Sciences 2003, 43 (1) , 36-44. https ...
[6]
[PDF] Section 13.1 Chemical Graph Theory - ResearchGate
In general, a graph is used to represent a molecule by considering the atoms as the vertices of the graph and the molecular bonds as the edges.
[7]
Graph theoretic and machine learning approaches in molecular ...
Jul 31, 2025 · Such graphs are typically undirected and simple, having no loop-edges or multiple edges between the same pair of vertices. Chemical graph theory ...
[8]
Molecular Graphs and Molecular Hypergraphs of Organic Compounds
A molecular graph is a weighted (or labeled) graph with vertices corresponding to atoms of molecule, and with edges corresponding to chemical bonds in it. The ...
[9]
On Strong Metric Dimension of the Methane Molecular Graph Using ...
This study focuses on evaluating Molecular graph's methane CH 4 strong metric dimension which is a fundamental molecule in organic chemistry.
[10]
Molecular representations in AI-driven drug discovery: a review and ...
Sep 17, 2020 · Molecular graphs are generally undirected, meaning that the pairs in E are unordered. [36]. To map a graph from an abstract mathematical ...<|control11|><|separator|>
[11]
SMILES, a chemical language and information system. 1 ...
SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules.
[12]
A standard method to generate canonical SMILES based on the InChI
Sep 18, 2012 · In order to generate a unique, or canonical, representation for a molecule, one needs a way to canonically label the atoms of the molecule. Such ...
[13]
Topological/chemical distances and graph centers in molecular ...
Abstract. A set of priority rules is defined for obtaining the graph center in multigraphs, i.e. in molecular graphs with multiple bonds.
[14]
The depleted hydrogen atoms in chemical graph theory
Aug 6, 2025 · Graph concepts can also be used to tackle the problem of the hydrogen contribution in hydrogen depleted graphs, which are encoded by the aid ...
[15]
The Vertex-Connectivity Index Revisited - ACS Publications
The vertex-degree d is equal to the number of adjacent vertexes in a given (molecular) graph G. Two vertexes in G are adjacent if there is an edge joining them.
[16]
Formal Charges in Organic Chemistry
What we notice is that carbon has four bonds in each molecule, oxygen has two, nitrogen–three and halogens, together with hydrogen tend to have only one bond.
[17]
Implicit, Explicit and Query Hydrogens - Chemaxon Docs
The implicit hydrogen count is calculated from the valence of the atom. It equals the valence of the atom minus the valence calculated from the bond connections ...
[18]
https://www.chemistrysteps.com/valency-formal-charges-organic-chemistry/
[19]
Chemical Graph Theory | Nenad Trinajstic - Taylor & Francis eBooks
May 11, 2018 · This unique book offers a basic introduction to the handling of molecular graphs - mathematical diagrams representing molecular structures.Missing: connectivity cycles rings
[20]
Group graph: a molecular graph representation with enhanced ...
Nov 28, 2024 · Molecular formulas depict how atoms and bonds build molecules, akin to how graphs use nodes and edges to map networks. In the atom graph, atoms ...
[21]
[PDF] NEW SPECTRAL INDICES FOR MOLECULE DESCRIPTION
Two families of molecular descriptors are proposed, based on new functions of the molecule spectrum derived from any molecular matrix M(w), calculated by ...
[22]
Spectral Moments of the Edge Adjacency Matrix in Molecular ...
In this work we examined spectral moments of the edge adjacency matrix of benzenoid hydrocarbons. We designed combinatorial formulas for spectral moments up to ...<|control11|><|separator|>
[23]
Lower bounds for the Zagreb indices of trees with given total ...
Oct 7, 2025 · The Zagreb indices provide insights into the degree distribution of a molecular graph, offering a connection to molecular stability and ...<|separator|>
[24]
Systematic benchmark of substructure search in molecular graphs
Jul 31, 2012 · In this paper, we present a systematic evaluation of Ullmann's and the VF2 subgraph isomorphism algorithms on molecular data.
[25]
Deep generative models for 3D molecular structure - ScienceDirect
Generating molecules directly in 3D bypasses conformation generation from molecular graphs, that can lead to a potentially infinite number of plausible ...
[26]
Inverse design of 3d molecular structures with conditional ... - Nature
Feb 21, 2022 · We propose a conditional generative neural network for 3d molecular structures with specified chemical and structural properties.
[27]
Efficient prediction of reaction paths through molecular graph ... - NIH
Here, we propose a novel approach to rapidly search reaction paths in a fully automated fashion by combining chemical theory and heuristics. A key idea of our ...
[28]
SynTemp: Efficient Extraction of Graph-Based Reaction Rules from ...
Chemical reactions can be modeled and studied through systems of rule-based rewriting of molecular graphs. (7,37,38) These graph transformation systems require ...
[29]
effectively measuring molecular diversity by shortest Hamiltonian ...
Aug 7, 2024 · We propose Hamiltonian diversity, a novel molecular diversity metric predicated upon the shortest Hamiltonian circuit.Missing: path | Show results with:path
[30]
Ligand-Based Virtual Screening Using Graph Edit Distance as ...
This study investigates the effectiveness of a graph-only driven molecular comparison by using extended reduced graphs along with graph edit distance methods.
[31]
On Topological Indices of Fractal and Cayley Tree Type Dendrimers
Aug 8, 2018 · Molecular graph is the illustration of a chemical compound or complex structure in which nodes correspond to atoms and edges correspond to ...
[32]
Neighborhood degree sum-based molecular descriptors of fractal ...
In this report, some neighborhood degree sum-based molecular descriptors are obtained for the fractal tree and the Cayley tree dendrimers.
[33]
QSPR Calculation of Normal Boiling Points of Organic Molecules ...
Dec 31, 2004 · The aim of this study is to present the results derived from the use of a particular sort of flexible molecular descriptors to estimate the BPs ...
[34]
Using machine learning in QSPR to estimate the boiling and critical ...
May 1, 2025 · This study aims to develop models using a Quantitative Structure-Property Relationship (QSPR) approach to predict Tb and Tc for 417 and 412 organic compounds, ...
[35]
Quantitative structure-property relationships for prediction of boiling ...
This paper reviews the QSPR prediction of boiling point, vapor pressure, and ... Use of electron–electron repulsion energy as a molecular descriptor in. QSAR and ...
[36]
Exploring Top Open-Source Toolkits for Cheminformatics: RDKit ...
The top open-source cheminformatics toolkits are RDKit, Open Babel, and CDK. RDKit is user-friendly, Open Babel converts structures, and CDK is for ...
[37]
Open Source Cheminformatics Toolkits - Macs in Chemistry
Feb 17, 2023 · The RDKit is an open source toolkit for cheminformatics, 2D and 3D molecular operations, descriptor generation for machine learning, etc.Openbabel (http://openbabel... · Rdkit (http://www.Rdkit.Org) · Chython (https://github...
[38]
Cheminformatics Microservice: unifying access to open ...
The Cheminformatics Microservice provides a unified interface for accessing RDKit, CDK, and Open Babel toolkits, and advanced functionalities.
[39]
Boosting RDKit molecular simulations through OpenMM | Cresset
Feb 20, 2017 · It features tight integration with the Jupyter notebook, so you can display your molecules in 2D and 3D interactively while you develop your ...
[40]
PDB2DAT: Automating LAMMPS data file generation from PDB ...
It integrates Rdkit and XYZ2MOL to generate molecular conformers, and Pysimm for assigning force field types and charges. •. Self-contained and lightweight, ...Original Software... · 1. Introduction · 2. Software Description
[41]
[1704.01212] Neural Message Passing for Quantum Chemistry - arXiv
Apr 4, 2017 · Neural Message Passing for Quantum Chemistry. Authors:Justin Gilmer, Samuel S. Schoenholz, Patrick F. Riley, Oriol Vinyals, George E. Dahl.
[42]
Leonhard Euler, “Solution of a problem in the geometry of position”
This famous paper on the bridges of Königsberg, in East Prussia, is generally considered to be the beginning of graph theory.
[43]
[PDF] On the mathematical theory of isomers - OEIS
Phil Mag Cayley 1874. 444. Prof. Cayley on the Mathematical Theory of Isomers. By addition of (A) and (B), a²=a²+6²₁₂. which, on writing x for x, is 18. We ...
[44]
Kombinatorische Anzahlbestimmungen für Gruppen, Graphen und ...
Download PDF ... Cite this article. Pólya, G. Kombinatorische Anzahlbestimmungen für Gruppen, Graphen und chemische Verbindungen. Acta Math. 68, 145–254 (1937).
[45]
The Dissimilarity Characteristic of Husimi Trees - jstor
1, July, 1953. Printed in U.S.A.. THE DISSIMILARITY CHARACTERISTIC OF HUSIMI TREES. By FRANK HARARY AND ROBERT Z. NORMAN. (Received May 12, 1952). (Revised July ...
[46]
[PDF] Alexandru T. Balaban
These graphs, published in 1966 [46] constitute the first reaction graphs, in which vertices symbolize reaction intermediates and edges symbolize elementary.
[47]
[PDF] Confessions and Reflections of a Graph-Theoretical Chemist, pp. 3-17
Topological indices (TIs) are numbers associated with chemical structures, i.e. with constitutional graphs. This is not a one-to-one association because ...
[48]
[PDF] History of QSAR - IMADA
The Wiener index increases with the number of atoms, and for a constant number of atoms, reaches a maximum for linear structures and a minimum for the most ...
[49]
The connectivity index 25 years after - ScienceDirect.com
We review the developments following introduction of the connectivity indices as molecular descriptors in multiple linear regression analysis (MLRA)
[50]
Daylight Theory: SMILES
SMILES denotes a molecular structure as a graph with optional chiral indications. This is essentially the two-dimensional picture chemists draw to describe a ...
[51]
Graphs in molecular biology | BMC Bioinformatics | Full Text
Sep 27, 2007 · Graph theoretical concepts are useful for the description and analysis of interactions and relationships in biological systems.
[52]
A Chemistry Recommendation Engine Built Using a Graph Database
Jul 15, 2017 · The Fragment Network is a graph database that allows a user to efficiently search chemical space around a compound of interest.