東華大學圖書館 |

Integrated in silico-in vitro Approaches to Toxicity Assessment and Molecular Design in 21st Century.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Integrated in silico-in vitro Approaches to Toxicity Assessment and Molecular Design in 21st Century./
作者:	Melnikov, Fjodor.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2019,
面頁冊數:	257 p.
附註:	Source: Dissertations Abstracts International, Volume: 81-10, Section: B.
Contained By:	Dissertations Abstracts International81-10B.
標題:	Toxicology. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=13905020
ISBN:	9798607312923

Integrated in silico-in vitro Approaches to Toxicity Assessment and Molecular Design in 21st Century.
Melnikov, Fjodor.

Integrated in silico-in vitro Approaches to Toxicity Assessment and Molecular Design in 21st Century. - Ann Arbor : ProQuest Dissertations & Theses, 2019 - 257 p.

Source: Dissertations Abstracts International, Volume: 81-10, Section: B.

Thesis (Ph.D.)--Yale University, 2019.

This item must not be sold to any third party vendors.

Natural and man-made molecules are widespread and essential components or modern industry and commerce. While some of these substances have received significant attention for their environmental and health hazards, most lack sufficient safety data for adequate hazard assessment. Furthermore, the controversial hazardous substances removed under social or political pressure are often replaced by molecules without safety data, resulting in regrettable substitutions. Consequently, it is imperative to develop new methods to assess chemical safety and to design functional and benign alternatives to hazardous chemical in commerce, according to principle of green chemistry. The first two decades of the 21st century saw a rapid shift in hazard and risk assessment away from expensive animal-based models. Greater emphasis has been given to measuring molecular perturbations and the chemical interactions between commercial chemicals and their biological targets. Simultaneously, new technologies enabled high throughput, targeted measurements of cellular perturbation in vitro, producing large volumes of new data on biological effects of chemicals. The new data streams and technologies present new opportunities for in silico methods in toxicity assessment and molecular design. However, with the new opportunities come new challenges. Large HTS data repositories require improved method of data inference to avoid erroneous conclusions. New types of in silico models are necessary to advance mechanistic relevance of outcomes measured in vitro to human health. Finally, in silico models for toxicity predictions require robust input properties, meaningful applicability domain definitions, sufficient transparence, and external validation to be useful in regulatory assessment and rational molecular design. Ideally, the in silico hazard and design models should explicitly consider human intuition in the model structure and avoid purely associative relationships.The work covered by this dissertation focuses on improved inference from HTS data bases and developing mechanistically robust in silico tools for chemical assessment and design. Using the U.S. Tox21 in vitro toxicity database as an example, we evaluated likelihood-based statistical methods for their performance in qHTS inference. The analyses demonstrated that robust loss functions and error propagation techniques are necessary to analyze qHTS data, which is often contaminated with technical artifacts and random outliers (Chapter 2). Furthermore, detailed understanding of the technologies that produced the data sets is imperative when drawing conclusions from the data. In ratiometric assays, such as the ones used by Tox21 screening facility, it is crucial to consider multiple readouts to adequately infer chemical activity. In its turn the accurate activity inference is necessary to group chemicals by mechanism of action, develop molecular design guidelines, or assess chemical effects in vitro (Chapter 3).While many structure-activity models have been developed to predict chemical activity over the years, many of these tools lack mechanistic relevance to the target biological endpoint. In an assessment of several in silico tools for estimating acute aquatic toxicity, the models built on mechanistically-relevant chemical descriptors outperformed models based on associative relationships. Furthermore, the chemical escriptor estimation methods played a large role in model performance (Chapter 4). However, the assessed models focus on apical endpoints in model animals. The move towards perturbation-based models requires greater insights from a combination of testing methods, typically a series of in vitro studies. However, using several chemicals of regulatory concern, we showed that an integrated in vitro - in silico assessment can help elucidate the molecular perturbation associated with chemical toxicity. Furthermore, the integrate analysis can help identify molecular initiating events involved in the induction of oxidative stress (OS) by seven diverse chemicals. As shown previously, exposure to electrophilic chemicals at low concentrations deplete GSH initially and induce antioxidant defense system over time in Hepa-1. Computational studies quantitatively validated the contribution of chemical mechanisms and reactivity models to toxicity and GSH depletion. (Chapter 5). These analyses highlight the importance of data processing and the power of computational approaches for large-scale data inference and mechanistic toxicity assessment. We further highlight that no computation algorithm is perfect and must be accompanied by human insight. In chemistry and toxicology, the insights come from the theories of chemical behavior and chemical-biological interaction. Together robust in silico tools and mechanistic insights open new paths towards timely toxicity assessment, ration chemical design, substitution of hazardous chemicals with function and safe alternatives, and safer commercial products and processes.

ISBN: 9798607312923Subjects--Topical Terms:

556884
Toxicology.
Subjects--Index Terms:

Computational

Integrated in silico-in vitro Approaches to Toxicity Assessment and Molecular Design in 21st Century.
LDR:06258nmm a2200385 4500 001 2272663
005 20201105110150.5
008 220629s2019 ||||||||||||||||| ||eng d
020 $a 9798607312923
035 $a (MiAaPQ)AAI13905020
035 $a AAI13905020
040 $a MiAaPQ $c MiAaPQ
100 1 $a Melnikov, Fjodor. $3 3550092
245 1 0 $a Integrated in silico-in vitro Approaches to Toxicity Assessment and Molecular Design in 21st Century.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2019
300 $a 257 p.
500 $a Source: Dissertations Abstracts International, Volume: 81-10, Section: B.
500 $a Advisor: Anastas, Paul T;Zimmerman, Julie B.
502 $a Thesis (Ph.D.)--Yale University, 2019.
506 $a This item must not be sold to any third party vendors.
520 $a Natural and man-made molecules are widespread and essential components or modern industry and commerce. While some of these substances have received significant attention for their environmental and health hazards, most lack sufficient safety data for adequate hazard assessment. Furthermore, the controversial hazardous substances removed under social or political pressure are often replaced by molecules without safety data, resulting in regrettable substitutions. Consequently, it is imperative to develop new methods to assess chemical safety and to design functional and benign alternatives to hazardous chemical in commerce, according to principle of green chemistry. The first two decades of the 21st century saw a rapid shift in hazard and risk assessment away from expensive animal-based models. Greater emphasis has been given to measuring molecular perturbations and the chemical interactions between commercial chemicals and their biological targets. Simultaneously, new technologies enabled high throughput, targeted measurements of cellular perturbation in vitro, producing large volumes of new data on biological effects of chemicals. The new data streams and technologies present new opportunities for in silico methods in toxicity assessment and molecular design. However, with the new opportunities come new challenges. Large HTS data repositories require improved method of data inference to avoid erroneous conclusions. New types of in silico models are necessary to advance mechanistic relevance of outcomes measured in vitro to human health. Finally, in silico models for toxicity predictions require robust input properties, meaningful applicability domain definitions, sufficient transparence, and external validation to be useful in regulatory assessment and rational molecular design. Ideally, the in silico hazard and design models should explicitly consider human intuition in the model structure and avoid purely associative relationships.The work covered by this dissertation focuses on improved inference from HTS data bases and developing mechanistically robust in silico tools for chemical assessment and design. Using the U.S. Tox21 in vitro toxicity database as an example, we evaluated likelihood-based statistical methods for their performance in qHTS inference. The analyses demonstrated that robust loss functions and error propagation techniques are necessary to analyze qHTS data, which is often contaminated with technical artifacts and random outliers (Chapter 2). Furthermore, detailed understanding of the technologies that produced the data sets is imperative when drawing conclusions from the data. In ratiometric assays, such as the ones used by Tox21 screening facility, it is crucial to consider multiple readouts to adequately infer chemical activity. In its turn the accurate activity inference is necessary to group chemicals by mechanism of action, develop molecular design guidelines, or assess chemical effects in vitro (Chapter 3).While many structure-activity models have been developed to predict chemical activity over the years, many of these tools lack mechanistic relevance to the target biological endpoint. In an assessment of several in silico tools for estimating acute aquatic toxicity, the models built on mechanistically-relevant chemical descriptors outperformed models based on associative relationships. Furthermore, the chemical escriptor estimation methods played a large role in model performance (Chapter 4). However, the assessed models focus on apical endpoints in model animals. The move towards perturbation-based models requires greater insights from a combination of testing methods, typically a series of in vitro studies. However, using several chemicals of regulatory concern, we showed that an integrated in vitro - in silico assessment can help elucidate the molecular perturbation associated with chemical toxicity. Furthermore, the integrate analysis can help identify molecular initiating events involved in the induction of oxidative stress (OS) by seven diverse chemicals. As shown previously, exposure to electrophilic chemicals at low concentrations deplete GSH initially and induce antioxidant defense system over time in Hepa-1. Computational studies quantitatively validated the contribution of chemical mechanisms and reactivity models to toxicity and GSH depletion. (Chapter 5). These analyses highlight the importance of data processing and the power of computational approaches for large-scale data inference and mechanistic toxicity assessment. We further highlight that no computation algorithm is perfect and must be accompanied by human insight. In chemistry and toxicology, the insights come from the theories of chemical behavior and chemical-biological interaction. Together robust in silico tools and mechanistic insights open new paths towards timely toxicity assessment, ration chemical design, substitution of hazardous chemicals with function and safe alternatives, and safer commercial products and processes.
590 $a School code: 0265.
650 4 $a Toxicology. $3 556884
650 4 $a Computational chemistry. $3 3350019
650 4 $a Statistics. $3 517247
653 $a Computational
653 $a Green chemistry
653 $a In vitro-in silico
653 $a Molecular modeling
653 $a Small molecule design
653 $a Tox21
690 $a 0383
690 $a 0219
690 $a 0463
710 2 $a Yale University. $b Forestry and Environmental Studies. $3 3172516
773 0 $t Dissertations Abstracts International $g 81-10B.
790 $a 0265
791 $a Ph.D.
792 $a 2019
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=13905020