Graph Kernels for Molecular Structure−Activity Relationship Analysis with Support Vector Machines

Mahé, Pierre; Ueda, Nobuhisa; Akutsu, Tatsuya; Perret, Jean-Luc; Vert, Jean-Philippe

doi:10.1021/ci050039t

Cited by 161 publications

(144 citation statements)

References 21 publications

(72 reference statements)

Supporting

Mentioning

143

Contrasting

Order By: Relevance

“…This kernel is defined as the Tanimoto coefficient between fingerprints indicating the presence or absence of all possible molecular fragments of length up to 8 in the 2D structure of the molecule, where a fragment refers to a sequence of atoms connected by covalent bonds. We note that this fingerprint is similar to classical 2D-fingerprints such as the Daylight representation 4 , with the difference that our implementation does not require to fold the fingerprint into a small-size vector (18).…”

Section: Methodsmentioning

confidence: 92%

“…We describe this algorithm in the case of the three-points kernels (15), (16) and (17), its extension to the two-points kernels (18), (19) and (20) being straightfoward. Following the notation of Section 5, we represent molecules by complete, atom-based labeled graphs, with the difference that the set of atom labels L defining the vertices labels is considered to be discrete (e.g., the atom types), and the edges are now labeled by the bin index of the corresponding inter-atomic distance.…”

Section: Definition 8 (Two-points Tanimoto Kernel)mentioning

confidence: 99%

See 1 more Smart Citation

The Pharmacophore Kernel for Virtual Screening with Support Vector Machines

Mahé

Ralaivola

Stoven

et al. 2006

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

We introduce a family of positive definite kernels specifically optimized for the manipulation of 3D structures of molecules with kernel methods. The kernels are based on the comparison of the three-points pharmacophores present in the 3D structures of molecules, a set of molecular features known to be particularly relevant for virtual screening applications. We present a computationally demanding exact implementation of these kernels, as well as fast approximations related to the classical fingerprint-based approaches. Experimental results suggest that this new approach outperforms state-of-the-art algorithms based on the 2D structure of molecules for the detection of inhibitors of several drug targets.

show abstract

Section: Methodsmentioning

confidence: 92%

Section: Definition 8 (Two-points Tanimoto Kernel)mentioning

confidence: 99%

The Pharmacophore Kernel for Virtual Screening with Support Vector Machines

Mahé

Ralaivola

Stoven

et al. 2006

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Paths are allowed to self-intersect and traverse the same node twice, to capture ring structures, but are not allowed to traverse the same edge twice, to avoid "totters". 25 2.4. 2.5D Surface and 3D Kernels Based on Delaunay Tetrahedrizations.…”

Section: Kernel Methodsmentioning

confidence: 99%

One- to Four-Dimensional Kernels for Virtual Screening and the Prediction of Physical, Chemical, and Biological Properties

Azencott

Ksikes

Swamidass

et al. 2007

J. Chem. Inf. Model.

View full text Add to dashboard Cite

Many chemoinformatics applications, including high-throughput virtual screening, benefit from being able to rapidly predict the physical, chemical, and biological properties of small molecules to screen large repositories and identify suitable candidates. When training sets are available, machine learning methods provide an effective alternative to ab initio methods for these predictions. Here, we leverage rich molecular representations including 1D SMILES strings, 2D graphs of bonds, and 3D coordinates to derive efficient machine learning kernels to address regression problems. We further expand the library of available spectral kernels for small molecules developed for classification problems to include 2.5D surface and 3D kernels using Delaunay tetrahedrization and other techniques from computational geometry, 3D pharmacophore kernels, and 3.5D or 4D kernels capable of taking into account multiple molecular configurations, such as conformers. The kernels are comprehensively tested using cross-validation and redundancy-reduction methods on regression problems using several available data sets to predict boiling points, melting points, aqueous solubility, octanol/water partition coefficients, and biological activity with state-of-the art results. When sufficient training data are available, 2D spectral kernels in general tend to yield the best and most robust results, better than state-of-the art. On data sets containing thousands of molecules, the kernels achieve a squared correlation coefficient of 0.91 for aqueous solubility prediction and 0.94 for octanol/water partition coefficient prediction. Averaging over conformations improves the performance of kernels based on the three-dimensional structure of molecules, especially on challenging data sets. Kernel predictors for aqueous solubility (kSOL), LogP (kLOGP), and melting point (kMELT) are available over the Web through: http:// cdb.ics.uci.edu.

show abstract

“…When data are graphs, such as when we represent a molecular compound by its 2D planar representation, a possible kernel is based on the comparison of the walks in the graphs. [10,7,14] Another interesting class of kernels is the set of multitask kernels, useful when we want to fit different models by sharing information across the tasks. [6] More precisely, suppose we want to fit a model of the form f(x,y) where x is a cell line and y is a chemical, considered as a task in the sense that we want to fit a model specific to each chemical to predict its toxicity across cell lines.…”

Section: Kernels and Multitask Learningmentioning

confidence: 99%

Kernel Multitask Regression for Toxicogenetics

Bernard

Jiao

Scornet

et al. 2017

Molecular Informatics

Self Cite

View full text Add to dashboard Cite

The development of high-throughput in vitro assays to study quantitatively the toxicity of chemical compounds on genetically characterized human-derived cell lines paves the way to predictive toxicogenetics, where one would be able to predict the toxicity of any particular compound on any particular individual. In this paper we present a machine learning-based approach for that purpose, kernel multitask regression (KMR), which combines chemical characterizations of molecular compounds with genetic and transcriptomic characterizations of cell lines to predict the toxicity of a given compound on a given cell line. We demonstrate the relevance of the method on the recent DREAM8 Toxicogenetics challenge, where it ranked among the best state-of-the-art models, and discuss the importance of choosing good descriptors for cell lines and chemicals.

show abstract

Graph Kernels for Molecular Structure−Activity Relationship Analysis with Support Vector Machines

Cited by 161 publications

References 21 publications

The Pharmacophore Kernel for Virtual Screening with Support Vector Machines

The Pharmacophore Kernel for Virtual Screening with Support Vector Machines

One- to Four-Dimensional Kernels for Virtual Screening and the Prediction of Physical, Chemical, and Biological Properties

Kernel Multitask Regression for Toxicogenetics

Contact Info

Product

Resources

About