Cargando…

Neuraldecipher – reverse-engineering extended-connectivity fingerprints (ECFPs) to their molecular structures

Protecting molecular structures from disclosure against external parties is of great relevance for industrial and private associations, such as pharmaceutical companies. Within the framework of external collaborations, it is common to exchange datasets by encoding the molecular structures into descr...

Descripción completa

Detalles Bibliográficos
Autores principales: Le, Tuan, Winter, Robin, Noé, Frank, Clevert, Djork-Arné
Formato: Online Artículo Texto
Lenguaje:English
Publicado: The Royal Society of Chemistry 2020
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8162443/
https://www.ncbi.nlm.nih.gov/pubmed/34094299
http://dx.doi.org/10.1039/d0sc03115a
_version_ 1783700713886449664
author Le, Tuan
Winter, Robin
Noé, Frank
Clevert, Djork-Arné
author_facet Le, Tuan
Winter, Robin
Noé, Frank
Clevert, Djork-Arné
author_sort Le, Tuan
collection PubMed
description Protecting molecular structures from disclosure against external parties is of great relevance for industrial and private associations, such as pharmaceutical companies. Within the framework of external collaborations, it is common to exchange datasets by encoding the molecular structures into descriptors. Molecular fingerprints such as the extended-connectivity fingerprints (ECFPs) are frequently used for such an exchange, because they typically perform well on quantitative structure–activity relationship tasks. ECFPs are often considered to be non-invertible due to the way they are computed. In this paper, we present a fast reverse-engineering method to deduce the molecular structure given revealed ECFPs. Our method includes the Neuraldecipher, a neural network model that predicts a compact vector representation of compounds, given ECFPs. We then utilize another pre-trained model to retrieve the molecular structure as SMILES representation. We demonstrate that our method is able to reconstruct molecular structures to some extent, and improves, when ECFPs with larger fingerprint sizes are revealed. For example, given ECFP count vectors of length 4096, we are able to correctly deduce up to 69% of molecular structures on a validation set (112 K unique samples) with our method.
format Online
Article
Text
id pubmed-8162443
institution National Center for Biotechnology Information
language English
publishDate 2020
publisher The Royal Society of Chemistry
record_format MEDLINE/PubMed
spelling pubmed-81624432021-06-04 Neuraldecipher – reverse-engineering extended-connectivity fingerprints (ECFPs) to their molecular structures Le, Tuan Winter, Robin Noé, Frank Clevert, Djork-Arné Chem Sci Chemistry Protecting molecular structures from disclosure against external parties is of great relevance for industrial and private associations, such as pharmaceutical companies. Within the framework of external collaborations, it is common to exchange datasets by encoding the molecular structures into descriptors. Molecular fingerprints such as the extended-connectivity fingerprints (ECFPs) are frequently used for such an exchange, because they typically perform well on quantitative structure–activity relationship tasks. ECFPs are often considered to be non-invertible due to the way they are computed. In this paper, we present a fast reverse-engineering method to deduce the molecular structure given revealed ECFPs. Our method includes the Neuraldecipher, a neural network model that predicts a compact vector representation of compounds, given ECFPs. We then utilize another pre-trained model to retrieve the molecular structure as SMILES representation. We demonstrate that our method is able to reconstruct molecular structures to some extent, and improves, when ECFPs with larger fingerprint sizes are revealed. For example, given ECFP count vectors of length 4096, we are able to correctly deduce up to 69% of molecular structures on a validation set (112 K unique samples) with our method. The Royal Society of Chemistry 2020-09-11 /pmc/articles/PMC8162443/ /pubmed/34094299 http://dx.doi.org/10.1039/d0sc03115a Text en This journal is © The Royal Society of Chemistry https://creativecommons.org/licenses/by/3.0/
spellingShingle Chemistry
Le, Tuan
Winter, Robin
Noé, Frank
Clevert, Djork-Arné
Neuraldecipher – reverse-engineering extended-connectivity fingerprints (ECFPs) to their molecular structures
title Neuraldecipher – reverse-engineering extended-connectivity fingerprints (ECFPs) to their molecular structures
title_full Neuraldecipher – reverse-engineering extended-connectivity fingerprints (ECFPs) to their molecular structures
title_fullStr Neuraldecipher – reverse-engineering extended-connectivity fingerprints (ECFPs) to their molecular structures
title_full_unstemmed Neuraldecipher – reverse-engineering extended-connectivity fingerprints (ECFPs) to their molecular structures
title_short Neuraldecipher – reverse-engineering extended-connectivity fingerprints (ECFPs) to their molecular structures
title_sort neuraldecipher – reverse-engineering extended-connectivity fingerprints (ecfps) to their molecular structures
topic Chemistry
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8162443/
https://www.ncbi.nlm.nih.gov/pubmed/34094299
http://dx.doi.org/10.1039/d0sc03115a
work_keys_str_mv AT letuan neuraldecipherreverseengineeringextendedconnectivityfingerprintsecfpstotheirmolecularstructures
AT winterrobin neuraldecipherreverseengineeringextendedconnectivityfingerprintsecfpstotheirmolecularstructures
AT noefrank neuraldecipherreverseengineeringextendedconnectivityfingerprintsecfpstotheirmolecularstructures
AT clevertdjorkarne neuraldecipherreverseengineeringextendedconnectivityfingerprintsecfpstotheirmolecularstructures