Analysis and visualization of short time series data
TimeSeriesExperiment is a package for visualization and analysis of short,
regular time-series datasets. The package is a comprehensive toolbox built on
TimeSeriesExperiment class (extension of SummarizedExperiment) designed
specifically to handle time-course data. Functions included allow user to
perform the analysis efficiently. Apart from native functions,
TimeSeriesExperiment also includes wrappers around functions from other
package to let the user easily work on TimeSeriesExperiment object without
having to switch between different frameworks. The package is mostly indended
for gene expression data, but can be applied to any datasets with observations
taken over time.
TimeSeriesExperiment performs:
- data normalization and transformation
- heatmap plotting with colorbars
- PCA projection of samples
- PCA projection of features with corresponding time-series overlayed
- Time-series features clustering
- Finding genes differentially expressed between two conditions/groups at specific timepoints
- Finding genes with differential trajectories between two conditions/groups
- Pathway enrichment analysis. Looking for over-representation of differential genes in GO or KEGG pathways.
To install the package do the following:
install.packages("devtools")
devtools::install_github("nlhuong/TimeSeriesExperiment")
The package is currently submitted to Bioconductor and will be available in the future from:
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
BiocManager::install("TimeSeriesExperiment")
The examples in the package are based on the data from the study A comparative study of endoderm differentiation in humans and chimpanzees by Blake et al. (2018).
The data used in the vignettes on the Cop1 role in pro-inflammatory response can be accessed from GEO with accession number, GSE114762.