Dataset Information

Molecular graph convolutions: moving beyond fingerprints.

ABSTRACT: Molecular "fingerprints" encoding structural information are the workhorse of cheminformatics and machine learning in drug discovery applications. However, fingerprint representations necessarily emphasize particular aspects of the molecular structure while ignoring others, rather than allowing the model to make data-driven decisions. We describe molecular graph convolutions, a machine learning architecture for learning from undirected graphs, specifically small molecules. Graph convolutions use a simple encoding of the molecular graph-atoms, bonds, distances, etc.-which allows the model to take greater advantage of information in the graph structure. Although graph convolutions do not outperform all fingerprint-based methods, they (along with other graph-based methods) represent a new paradigm in ligand-based virtual screening with exciting opportunities for future improvement.

SUBMITTER: Kearnes S

PROVIDER: S-EPMC5028207 | biostudies-literature | 2016 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Molecular graph convolutions: moving beyond fingerprints.

Kearnes Steven S McCloskey Kevin K Berndl Marc M Pande Vijay V Riley Patrick P

Journal of computer-aided molecular design 20160824 8

Molecular "fingerprints" encoding structural information are the workhorse of cheminformatics and machine learning in drug discovery applications. However, fingerprint representations necessarily emphasize particular aspects of the molecular structure while ignoring others, rather than allowing the model to make data-driven decisions. We describe molecular graph convolutions, a machine learning architecture for learning from undirected graphs, specifically small molecules. Graph convolutions use ...[more]

PMID: 27558503

Similar Datasets

Project description:Livestock predation is a global problem and constitutes the main source of conflict between large carnivores and human interests. In Latin America, both jaguar and puma are known to prey on livestock, yet studies in Mesoamerica have been scattered and few have been carried out in Honduras. We interviewed ranchers in a biosphere reserve where jaguars and pumas are present. Local indigenous communities reported livestock predation (average annual loss of 7% from 2010-2019), with preventive and retaliatory killing as their main actions against predation by the jaguar and puma. Other sources of cattle loss included diseases and theft. The extensive management system (free grazing) lets cattle access forests where predators are more common. We found that livestock predation is not random, but rather, related to landscape variables and human influence. Sites farther from human influence and closer to forest cover were more susceptible to predation. Jaguar and puma persistence in the biosphere reserve will require measures that facilitate human-carnivore coexistence and comply with Sustainable Development Goals (SDG) 2 and 15 (zero hunger and biodiversity conservation). We propose management practices to mitigate livestock predation in the presence of large carnivores based on examples of proven human-carnivore coexistence in Venezuela, Brazil, Paraguay, and Nicaragua, such as improving the spatial arrangement of livestock (maintaining a distance from forest areas) and the incorporation of confinement pens for young calves (at least the first three months of life) and their mothers. If the pens are built close to the property's house and have constant surveillance and/or dogs, the results are likely to be more effective. Deploying these proven tools may help change the current negative perception of ranchers towards large carnivores that is essential to conservation under the aims of SDG 15. We recommend government policies and support aimed to strengthen livestock health to increase productivity and to reduce their vulnerability to predation. Finally, this study represents a baseline to understand the magnitude of the human-carnivore conflict over cattle in one of the largest biosphere reserves in Mesoamerica.

Project description:BACKGROUND:Interaction fingerprints (IFP) have been repeatedly shown to be valuable tools in virtual screening to identify novel hit compounds that can subsequently be optimized to drug candidates. As a complementary method to ligand docking, IFPs can be applied to quantify the similarity of predicted binding poses to a reference binding pose. For this purpose, a large number of similarity metrics can be applied, and various parameters of the IFPs themselves can be customized. In a large-scale comparison, we have assessed the effect of similarity metrics and IFP configurations to a number of virtual screening scenarios with ten different protein targets and thousands of molecules. Particularly, the effect of considering general interaction definitions (such as Any Contact, Backbone Interaction and Sidechain Interaction), the effect of filtering methods and the different groups of similarity metrics were studied. RESULTS:The performances were primarily compared based on AUC values, but we have also used the original similarity data for the comparison of similarity metrics with several statistical tests and the novel, robust sum of ranking differences (SRD) algorithm. With SRD, we can evaluate the consistency (or concordance) of the various similarity metrics to an ideal reference metric, which is provided by data fusion from the existing metrics. Different aspects of IFP configurations and similarity metrics were examined based on SRD values with analysis of variance (ANOVA) tests. CONCLUSION:A general approach is provided that can be applied for the reliable interpretation and usage of similarity measures with interaction fingerprints. Metrics that are viable alternatives to the commonly used Tanimoto coefficient were identified based on a comparison with an ideal reference metric (consensus). A careful selection of the applied bits (interaction definitions) and IFP filtering rules can improve the results of virtual screening (in terms of their agreement with the consensus metric). The open-source Python package FPKit was introduced for the similarity calculations and IFP filtering; it is available at: https://github.com/davidbajusz/fpkit .

Dataset Information

Molecular graph convolutions: moving beyond fingerprints.

Publications

Molecular graph convolutions: moving beyond fingerprints.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets