Grantee Research Project Results
2018 Progress Report: A Pipeline for in vitro-to-in vivo Extrapolation, Population Modeling, & Prioritization
EPA Grant Number: R835802C003Subproject: this is subproject number 003 , established and managed by the Center Director under grant R835802
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
Center: UT Center for Infrastructure Modeling and Management
Center Director: Hodges, Ben R.
Title: A Pipeline for in vitro-to-in vivo Extrapolation, Population Modeling, & Prioritization
Investigators: Wright, Fred A. , Wetmore, Barbara , Reif, David , Zhou, Yihui
Institution: North Carolina State University
EPA Project Officer: Aja, Hayley
Project Period: June 1, 2015 through May 31, 2019 (Extended to May 31, 2020)
Project Period Covered by this Report: June 1, 2018 through May 31,2019
RFA: Organotypic Culture Models for Predictive Toxicology Center (2013) RFA Text | Recipients Lists
Research Category: Chemical Safety for Sustainability
Objective:
The activities in Projects 1 and 2 necessitate a targeted, yet comprehensive analytic pipeline for analyzing organotypic culture model data screened for dosimetry, physiological parameters, and genetic and transcriptomic profiles, culminating in an informed basis for ranking and prioritization. Project 3 is responsible to collate, analyze, and synthesize the results from Projects 1 and 2. The data gathered in Projects 1 and 2 includes numerous screening measurements in human iPSC-derived cardiomyocytes and rodent embryonic stem cells, building upon recent advances to maximize the informativeness of beating cardiomyocyte models. The activities in the Project have followed the original proposal, with screening methods focusing on concentration-response modeling and multivariate analysis. By harnessing these efforts in a coherent pipeline, the Project synergizes with Projects 1 and 2 and adds value to the entire Center activity. In this year the Project has largely reached its goal of developing a coordinated analysis and decision-support pipeline based on complex data from an organotypic culture model systems in humans and mice, resulting in standard approaches and tools that can be used in future cardiotoxicity screening and inform human health assessments. Efforts in the upcoming year will be devoted to finalizing the pipelines and methods. The research goals are being achieved by pursuing the following specific objectives.
Specific Objective 1: To apply and refine methods to use the pharmacokinetic data from Project 1 to perform in vitro-to-in vivo extrapolation and subsequent generation of oral equivalent doses. [This Objective has been moved to Project 1 at Texas A&M University, as described in detail in previous reports]
Specific Objective 2: To apply concentration-response modeling to establish robust and appropriate points of departure.
Specific Objective 3: To perform annotation-informed analyses of population variation and association.
Specific Objective 4: To perform ranking and prioritization analyses of the ToxCast chemicals screened.
Progress Summary:
As was the case last year, most of the work in the past year has focused on the activities under proposed Specific Objectives 2, 3, and 4. To date, we have accomplished the following:
- We have refined our published improved pipeline for dose-response analysis, applicable toboth physiological parameters and to expression data (e.g. the TempOSeq technology). Thepipeline includes methods for sequence read counting and numerous flags in order to highlight genes that show evidence of differential expression prior to dose-response analysis. We have had numerous discussion with stakeholders at the U.S. EPA, NIEHS, and other environmental scientists about the implications of our work and high-throughput dose-response modeling.
- Downsampling analyses of the TempoSeq pipeline has been revealing in better understanding the role of replication in high-throughput dose-response modeling, when the total number of assays is held fixed. Some of the results can be counter-intuitive compared to standard toxicological design practice. E.g. the lack of replication can enable exploration of additional doses, etc., so that replication may not strictly be necessary.
- We have made considerable progress in scripting all analyses using R/Markdown, which enhances reproducible research and makes interaction with non-data scientists much more accessible, and easier to make modifications "on the fly."
- We have created expression analysis apps using R/Shiny, enabling fast and transparent analyses of high-dimensional datasets and contrast plots easier to run and display.
- In ToxPi 2.0, we have made publicly available finalized modifications of the software to
- handle multiple dimensions of prioritization
- to handle missing data, and
- quantify across endpoints with irregular correlation structure.
- more easily share results across collaborators.
The new toxpi software has been deployed at http://www.toxpi.org/ and is quickly developing a user base.
As discussed in previous reports, the ability to cluster samples based on the ToxPi profile, rather than an overall sum of slice values (the ToxPi rank) offers new insights into the data and has been received well by the toxicological community. These features are also already being used in analyses of Projects 1 and 2.
- Following on the ToxPi clustering methods, we have continued to make progress in evaluating clustering trees in a manner that enables comparison of different trees to each other, and to proposed prior categorization of chemicals. These methods work well with measures of toxicity in profiling of cardiomyocytes, and are important for other Project activity.
Future Activities:
- Bringing the published TempOSeq dose-response analysis pipeline to a new version, with automatic statistical testing and confidence intervals for Hill parameters, and downsampling analyses to understand the robustness to sample size
- Continue to analyze the data from high-content screening and high-throughput transcriptomics, working closely with Project 1 and 2 personnel. These involve standard statistical methods and tools applied to the cardiomyocyte data.
- Integrate information from additional data streams, including high-content screening platforms for neurotoxicological activity, public toxicological databases, and epidemiological studies.
- Exploring the impact of clustering techniques into ToxPi, so that best principles and standards can be applied and assessed visualized. Various stakeholders will be engaged in assessing these improvements, with suggestions for further efforts.
- Implementing the developed protocol for variability analysis across cell lines. These will be used in concert with other developing chemical prioritization schemes, e.g. in ToxPi, and perhaps using these measures as "slices" in ToxPi analyses.
Journal Articles on this Report : 6 Displayed | Download in RIS Format
Other subproject views: | All 18 publications | 10 publications in selected types | All 10 journal articles |
---|---|---|---|
Other center views: | All 150 publications | 45 publications in selected types | All 45 journal articles |
Type | Citation | ||
---|---|---|---|
|
Chiu WA, Guyton KZ, Martin MT, Reif DM, Rusyn I. Use of high-throughput in vitro toxicity screening data in cancer hazard evaluations by IARC Monograph Working Groups. ALTEX 2018;35(1):51-64. |
R835802 (2017) R835802 (2018) R835802C003 (2018) |
Exit Exit |
|
Grimm FA, Blanchette A, House JS, Ferguson K, Hsieh NH, Dalaijamts C, Wright AA, Anson B, Wright FA, Chiu WA, Rusyn I. A human population-based organotypic in vitro model for cardiotoxicity screening. ALTEX 2018;35:441-452. |
R835802 (2018) R835802C001 (2018) R835802C003 (2018) |
Exit Exit |
|
Marvel SW, To K, Grimm FA, Wright FA, Rusyn I, Reif DM. ToxPi Graphical User Interface 2.0: dynamic exploration, visualization, and sharing of integrated data models. BMC Bioinformatics 2018;19(1):80 (7 pp.). |
R835802 (2017) R835802 (2018) R835802C001 (2018) R835802C003 (2018) |
Exit Exit Exit |
|
Li G, Jima D, Wright FA, Nobel AB. HT-eQTL:integrative expression quantitative trait loci analysis in a large number of human tissues. BMC Bioinformatics 2018;19:95. |
R835802 (2018) R835802C003 (2018) |
|
|
Kosnik MB, Reif DM. Determination of chemical-disease risk values to prioritize connections between environmental factors, genetic variants, and human diseases. Toxicology and Applied Pharmacology2019;379:114674. |
R835802C003 (2018) |
Exit Exit |
|
Kosnik MB, Planchart A, Marvel SW, Reif DM, Mattingly CJ. Integration of curated and high-throughput screening data to elucidate environmental influences on disease pathways. Computational Toxicology2019;12:100094. |
R835802C003 (2018) |
Exit Exit |
Supplemental Keywords:
cardiovascular, stem cells, toxicity pathway, variability, dose-responseRelevant Websites:
ToxPi : Toxicological Prioritization Index Exit
Progress and Final Reports:
Original AbstractMain Center Abstract and Reports:
R835802 UT Center for Infrastructure Modeling and Management Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
R835802C001 High-throughput Hazard,Dose-responseandPopulationVariabilityAssessmentofCardiotoxicity in aHumanInducedPluripotentStem Cell(iPSC)-derivedinvitro Culture Model
R835802C002 Linking in vitro-to-in vivoToxicity Testing Using
Genetically-matchedOrganoids and Mice from a Novel Genetic Reference Population
R835802C003 A Pipeline for in vitro-to-in vivo Extrapolation, Population Modeling,
& Prioritization
The perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.
Project Research Results
10 journal articles for this subproject
Main Center: R835802
150 publications for this center
45 journal articles for this center