Naïve Bayesian Models for Vero Cell Cytotoxicity

Alexander L. Perryman, Jimmy S. Patel, Riccardo Russo, Eric Singleton, Nancy Connell, Sean Ekins, Joel S. Freundlich

Research output: Contribution to journalArticle

Abstract

Purpose: To advance translational research of potential therapeutic small molecules against infectious microbes, the compounds must display a relative lack of mammalian cell cytotoxicity. Vero cell cytotoxicity (CC50) is a common initial assay for this metric. We explored the development of naïve Bayesian models that can enhance the probability of identifying non-cytotoxic compounds. Methods: Vero cell cytotoxicity assays were identified in PubChem, reformatted, and curated to create a training set with 8741 unique small molecules. These data were used to develop Bayesian classifiers, which were assessed with internal cross-validation, external tests with a set of 193 compounds from our laboratory, and independent validation with an additional diverse set of 1609 unique compounds from PubChem. Results: Evaluation with independent, external test and validation sets indicated that cytotoxicity Bayesian models constructed with the ECFP_6 descriptor were more accurate than those that used FCFP_6 fingerprints. The best cytotoxicity Bayesian model displayed predictive power in external evaluations, according to conventional and chance-corrected statistics, as well as enrichment factors. Conclusions: The results from external tests demonstrate that our novel cytotoxicity Bayesian model displays sufficient predictive power to help guide translational research. To assist the chemical tool and drug discovery communities, our curated training set is being distributed as part of the Supplementary Material. [Figure not available: see fulltext.].

Original languageEnglish (US)
Article number170
JournalPharmaceutical Research
Volume35
Issue number9
DOIs
StatePublished - Sep 1 2018
Externally publishedYes

Fingerprint

Vero Cells
Translational Medical Research
Cytotoxicity
Dermatoglyphics
Drug Discovery
Assays
Molecules
Classifiers
Therapeutics
Cells
Statistics

Keywords

  • Bayesian model
  • machine learning
  • predicting mammalian cytotoxicity
  • translational research
  • vero cell CC

ASJC Scopus subject areas

  • Biotechnology
  • Molecular Medicine
  • Pharmacology
  • Pharmaceutical Science
  • Organic Chemistry
  • Pharmacology (medical)

Cite this

Perryman, A. L., Patel, J. S., Russo, R., Singleton, E., Connell, N., Ekins, S., & Freundlich, J. S. (2018). Naïve Bayesian Models for Vero Cell Cytotoxicity. Pharmaceutical Research, 35(9), [170]. https://doi.org/10.1007/s11095-018-2439-9

Naïve Bayesian Models for Vero Cell Cytotoxicity. / Perryman, Alexander L.; Patel, Jimmy S.; Russo, Riccardo; Singleton, Eric; Connell, Nancy; Ekins, Sean; Freundlich, Joel S.

In: Pharmaceutical Research, Vol. 35, No. 9, 170, 01.09.2018.

Research output: Contribution to journalArticle

Perryman, AL, Patel, JS, Russo, R, Singleton, E, Connell, N, Ekins, S & Freundlich, JS 2018, 'Naïve Bayesian Models for Vero Cell Cytotoxicity', Pharmaceutical Research, vol. 35, no. 9, 170. https://doi.org/10.1007/s11095-018-2439-9
Perryman AL, Patel JS, Russo R, Singleton E, Connell N, Ekins S et al. Naïve Bayesian Models for Vero Cell Cytotoxicity. Pharmaceutical Research. 2018 Sep 1;35(9). 170. https://doi.org/10.1007/s11095-018-2439-9
Perryman, Alexander L. ; Patel, Jimmy S. ; Russo, Riccardo ; Singleton, Eric ; Connell, Nancy ; Ekins, Sean ; Freundlich, Joel S. / Naïve Bayesian Models for Vero Cell Cytotoxicity. In: Pharmaceutical Research. 2018 ; Vol. 35, No. 9.
@article{61612adaf0eb4fdd95972ff7b017d433,
title = "Na{\"i}ve Bayesian Models for Vero Cell Cytotoxicity",
abstract = "Purpose: To advance translational research of potential therapeutic small molecules against infectious microbes, the compounds must display a relative lack of mammalian cell cytotoxicity. Vero cell cytotoxicity (CC50) is a common initial assay for this metric. We explored the development of na{\"i}ve Bayesian models that can enhance the probability of identifying non-cytotoxic compounds. Methods: Vero cell cytotoxicity assays were identified in PubChem, reformatted, and curated to create a training set with 8741 unique small molecules. These data were used to develop Bayesian classifiers, which were assessed with internal cross-validation, external tests with a set of 193 compounds from our laboratory, and independent validation with an additional diverse set of 1609 unique compounds from PubChem. Results: Evaluation with independent, external test and validation sets indicated that cytotoxicity Bayesian models constructed with the ECFP_6 descriptor were more accurate than those that used FCFP_6 fingerprints. The best cytotoxicity Bayesian model displayed predictive power in external evaluations, according to conventional and chance-corrected statistics, as well as enrichment factors. Conclusions: The results from external tests demonstrate that our novel cytotoxicity Bayesian model displays sufficient predictive power to help guide translational research. To assist the chemical tool and drug discovery communities, our curated training set is being distributed as part of the Supplementary Material. [Figure not available: see fulltext.].",
keywords = "Bayesian model, machine learning, predicting mammalian cytotoxicity, translational research, vero cell CC",
author = "Perryman, {Alexander L.} and Patel, {Jimmy S.} and Riccardo Russo and Eric Singleton and Nancy Connell and Sean Ekins and Freundlich, {Joel S.}",
year = "2018",
month = "9",
day = "1",
doi = "10.1007/s11095-018-2439-9",
language = "English (US)",
volume = "35",
journal = "Pharmaceutical Research",
issn = "0724-8741",
publisher = "Springer New York",
number = "9",

}

TY - JOUR

T1 - Naïve Bayesian Models for Vero Cell Cytotoxicity

AU - Perryman, Alexander L.

AU - Patel, Jimmy S.

AU - Russo, Riccardo

AU - Singleton, Eric

AU - Connell, Nancy

AU - Ekins, Sean

AU - Freundlich, Joel S.

PY - 2018/9/1

Y1 - 2018/9/1

N2 - Purpose: To advance translational research of potential therapeutic small molecules against infectious microbes, the compounds must display a relative lack of mammalian cell cytotoxicity. Vero cell cytotoxicity (CC50) is a common initial assay for this metric. We explored the development of naïve Bayesian models that can enhance the probability of identifying non-cytotoxic compounds. Methods: Vero cell cytotoxicity assays were identified in PubChem, reformatted, and curated to create a training set with 8741 unique small molecules. These data were used to develop Bayesian classifiers, which were assessed with internal cross-validation, external tests with a set of 193 compounds from our laboratory, and independent validation with an additional diverse set of 1609 unique compounds from PubChem. Results: Evaluation with independent, external test and validation sets indicated that cytotoxicity Bayesian models constructed with the ECFP_6 descriptor were more accurate than those that used FCFP_6 fingerprints. The best cytotoxicity Bayesian model displayed predictive power in external evaluations, according to conventional and chance-corrected statistics, as well as enrichment factors. Conclusions: The results from external tests demonstrate that our novel cytotoxicity Bayesian model displays sufficient predictive power to help guide translational research. To assist the chemical tool and drug discovery communities, our curated training set is being distributed as part of the Supplementary Material. [Figure not available: see fulltext.].

AB - Purpose: To advance translational research of potential therapeutic small molecules against infectious microbes, the compounds must display a relative lack of mammalian cell cytotoxicity. Vero cell cytotoxicity (CC50) is a common initial assay for this metric. We explored the development of naïve Bayesian models that can enhance the probability of identifying non-cytotoxic compounds. Methods: Vero cell cytotoxicity assays were identified in PubChem, reformatted, and curated to create a training set with 8741 unique small molecules. These data were used to develop Bayesian classifiers, which were assessed with internal cross-validation, external tests with a set of 193 compounds from our laboratory, and independent validation with an additional diverse set of 1609 unique compounds from PubChem. Results: Evaluation with independent, external test and validation sets indicated that cytotoxicity Bayesian models constructed with the ECFP_6 descriptor were more accurate than those that used FCFP_6 fingerprints. The best cytotoxicity Bayesian model displayed predictive power in external evaluations, according to conventional and chance-corrected statistics, as well as enrichment factors. Conclusions: The results from external tests demonstrate that our novel cytotoxicity Bayesian model displays sufficient predictive power to help guide translational research. To assist the chemical tool and drug discovery communities, our curated training set is being distributed as part of the Supplementary Material. [Figure not available: see fulltext.].

KW - Bayesian model

KW - machine learning

KW - predicting mammalian cytotoxicity

KW - translational research

KW - vero cell CC

UR - http://www.scopus.com/inward/record.url?scp=85049157944&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85049157944&partnerID=8YFLogxK

U2 - 10.1007/s11095-018-2439-9

DO - 10.1007/s11095-018-2439-9

M3 - Article

C2 - 29959603

AN - SCOPUS:85049157944

VL - 35

JO - Pharmaceutical Research

JF - Pharmaceutical Research

SN - 0724-8741

IS - 9

M1 - 170

ER -