es505199f_si_001.pdf (2.42 MB)
Toxicity Mechanisms Identification via Gene Set Enrichment Analysis of Time-Series Toxicogenomics Data: Impact of Time and Concentration
journal contribution
posted on 2015-04-07, 00:00 authored by Ce Gao, David Weisman, Jiaqi Lan, Na Gou, April Z. GuThe
advance in high-throughput “toxicogenomics” technologies,
which allows for concurrent monitoring of cellular responses globally
upon exposure to chemical toxicants, presents promises for next-generation
toxicity assessment. It is recognized that cellular responses to toxicants
have a highly dynamic nature, and exhibit both temporal complexity
and dose-response shifts. Most current gene enrichment or pathway
analysis lack the recognition of the inherent correlation within time
series data, and may potentially miss important pathways or yield
biased and inconsistent results that ignore dynamic patterns and time-sensitivity.
In this study, we investigated the application of two score metrics
for GSEA (gene set enrichment analysis) to rank the genes that consider
the temporal gene expression profile. One applies a novel time series
CPCA (common principal components analysis) to generate scores for
genes based on their contributions to the common temporal variation
among treatments for a given chemical at different concentrations.
Another one employs an integrated altered gene expression quantifier-TELI
(transcriptional effect level index) that integrates altered gene
expression magnitude over the exposure time. By comparing the GSEA
results using two different ranking metrics for examining the dynamic
responses of reporter cells treated with various dose levels of three
model toxicants, mitomycin C, hydrogen peroxide, and lead nitrate,
the analysis identified and revealed different toxicity mechanisms
of these chemicals that exhibit chemical-specific, as well as time-aware
and dose-sensitive nature. The ability, advantages, and disadvantages
of varying ranking metrics were discussed. These findings support
the notion that toxicity bioassays should account for the cells’
complex dynamic responses, thereby implying that both data acquisition
and data analysis should look beyond simple traditional end point
responses.
History
Usage metrics
Categories
Keywords
reporter cellsdata acquisitionhydrogen peroxidemitomycin Ccomponents analysisGSEA resultsexposure timedose levelsscore metricstime series dataenrichment analysistoxicity mechanismsfindings supportToxicity Mechanisms Identificationdata analysischemical toxicantsgene enrichmentnovel time series CPCAend point responsestoxicity bioassaystranscriptional effect level indexEnrichment Analysisgene expression magnitudepathway analysis lackConcentrationThe advancegene expression profilemodel toxicants
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC