Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database GDB-17

By Lars Ruddigkeit, Ruud Van Deursen, Lorenz C. Blum, Jean-Louis Reymond. 2012

1,435 citations recorded by the source corpus.

Drug molecules consist of a few tens of atoms connected by covalent bonds. How many such molecules are possible in total and what is their structure? This question is of pressing interest in medicinal chemistry to help solve the problems of drug potency, selectivity, and toxicity and reduce attrition rates by pointing to new molecular series. To better define the unknown chemical space, we have enumerated 166.4 billion molecules of up to 17 atoms of C, N, O, S, and halogens forming the chemical universe database GDB-17, covering a size range containing many drugs and typical for lead compounds. GDB-17 contains millions of isomers of known drugs, including analogs with high shape similarity to the parent drug. Compared to known molecules in PubChem, GDB-17 molecules are much richer in nonaromatic heterocycles, quaternary centers, and stereoisomers, densely populate the third dimension in shape space, and represent many more scaffold types.

Subfields: Computational Theory and Mathematics, Molecular Biology, Organic Chemistry, Evaluation & Benchmarks, Computational Chemistry & Drug Discovery

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Molecular Biology

In Organic Chemistry

In Evaluation & Benchmarks

In Computational Chemistry & Drug Discovery

Citations in the corpus

Neural Message Passing for Quantum Chemistry (2017, background): “The representation used in Rupp et al. (2012) is not invariant to graph isomorphism.”
Relevance of Rotationally Equivariant Convolutions for Predicting Molecular Properties (2020, positive): “The question considered in this paper can loosely be stated as: When doing regression on molecular properties, what is missing when one employs only invariant layers in a neural network as opposed to including equivariant ones? We explore this question using the QM9 benchmark [9, 10] by predicting q”
Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks (2018, methodology): “To demonstrate that our architectures scale to larger datasets and offer beneﬁts on real-world applications, we conduct experiments on the Q M 9 dataset (Ramakrishnan et al. 2014; Ruddigkeit et al. 2012; Wu et al. 2018), which consists of Target Method 133 385 small molecules.”
Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials (2018, background): “Quantum Machines 9 (QM9) (Ramakrishnan et al., 2014; Ruddigkeit et al., 2012) The dataset contains 133 885 examples of stable small organic molecules with up to 7 heavy atoms (CONF) and The Materials Project (Jain et al., 2013) This dataset contains geometries and formation energies of 69 640…”
Constrained Graph Variational Autoencoders for Molecule Design (2018, methodology): “On average, these molecules are bigger ( ∼ 23 heavy atoms) and structurally more complex than the molecules in QM9.”
MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Design (2022, background): “QM9 [193], [194] provides quantum chemical properties for a relevant, consistent, and comprehensive chemical space of small organic molecules.”
Inverse design of 3d molecular structures with conditional generative neural networks (2021, methodology): “The QM9 data set was assembled by relaxing structures from the GDB enumeration [61] of graphs for small organic compounds.”
Artificial intelligence to deep learning: machine intelligence approach for drug discovery (2021, background): “Further, it is a huge task to find a new drug when there are ~ 10 60 existing drug-like molecules [446].”