How thelovelab/deseq2 works
DESeq2 is an R package for differential analysis of count data, principally RNA-seq. Users provide counts and sample metadata, construct a DESeqDataSet, run the DESeq analysis workflow, and inspect statistical results and visualizations. The workflow includes normalization, dispersion estimation, and negative-binomial GLM fitting; Wald and likelihood-ratio tests produce results that can be further processed with fold-change shrinkage. The package also supports variance-stabilizing and regularized-log transformations. The README is empty and source excerpts are limited, so links for documented but unsampled capabilities are kept at the module level.
Main entry points: core.R DESeq Workflow · AllClasses.R DESeqDataSet · results.R Result Extraction · plots.R Plots
Components 16
Data and Workflow
- DESeqDataSet
R/AllClasses.R - DESeq Workflow
R/core.R
Model Fitting
- Size Factors
R/core.R - Dispersion Estimation
R/core.R - Negative Binomial GLMs
R/fitNbinomGLMs.R - Numerical Fitting
src/DESeq2.cpp
Inference
- Wald Test
R/core.R - Likelihood Ratio Test
R/core.R - Result Extraction
R/results.R - Fold Change Shrinkage
R/lfcShrink.R
Transform and Explore
- Variance Stabilization
R/vst.R - Regularized Log
R/rlog.R - Plots
R/plots.R - Count Transforms
R/wrappers.R
Other components
- Analyst
- Counts and Metadata
Connections 23
- Analyst → Counts and Metadata: provides
- Counts and Metadata → DESeqDataSet: constructs
- Analyst → DESeq Workflow: runs
- DESeqDataSet → DESeq Workflow: analyzed by
- DESeq Workflow → Size Factors: estimates
- DESeq Workflow → Dispersion Estimation: estimates
- DESeq Workflow → Negative Binomial GLMs: fits
- Dispersion Estimation → Numerical Fitting: calls
- Negative Binomial GLMs → Numerical Fitting: calls
- DESeq Workflow → Wald Test: runs
- DESeq Workflow → Likelihood Ratio Test: runs
- Wald Test → Result Extraction: produces
- Likelihood Ratio Test → Result Extraction: produces
- Result Extraction → Fold Change Shrinkage: refines
- Analyst → Variance Stabilization: requests
- Analyst → Regularized Log: requests
- DESeqDataSet → Variance Stabilization: transforms
- DESeqDataSet → Regularized Log: transforms
- DESeqDataSet → Count Transforms: transforms
- Analyst → Plots: requests
- Result Extraction → Plots: visualizes
- Variance Stabilization → Plots: visualizes
- Regularized Log → Plots: visualizes