epigen/dea_limma

A Snakemake workflow for performing and visualizing differential expression analyses (DEA) on NGS data powered by the R package limma.

Differential Analysis & Visualization Snakemake Workflow Using limma

A Snakemake workflow for performing and visualizing differential expression (or accessibility) analyses (DEA) of NGS data (eg RNA-seq, ATAC-seq, scRNA-seq,...) powered by the R package limma.

This workflow adheres to the module specifications of MR.PARETO, an effort to augment research by modularizing (biomedical) data science. For more details and modules check out the project's repository.

If you use this workflow in a publication, please don't forget to give credits to the authors by citing it using this DOI 10.5281/zenodo.7808516.

Table of contents

• Authors
• Software
• Methods
• Features
• Usage
• Configuration
• Examples
• Links
• Resources
• Publications

Authors

• Stephan Reichl
• Lukas Folkman
• Christoph Bock

Software

This project wouldn't be possible without the following software and their dependencies:

Software	Reference (DOI)
edgeR	https://doi.org/10.1093/bioinformatics/btp616
EnhancedVolcano	https://doi.org/10.18129/B9.bioc.EnhancedVolcano
ggplot2	https://ggplot2.tidyverse.org/
patchwork	https://CRAN.R-project.org/package=patchwork
pheatmap	https://cran.r-project.org/package=pheatmap
limma	https://doi.org/10.1093/nar/gkv007
Snakemake	https://doi.org/10.12688/f1000research.29032.2

Methods

This is a template for the Methods section of a scientific publication and is intended to serve as a starting point. Only retain paragraphs relevant to your analysis. References [ref] to the respective publications are curated in the software table above. Versions (ver) have to be read out from the respective conda environment specifications (.yaml file) or post execution. Parameters that have to be adapted depending on the data or workflow configurations are denoted in squared brackets e.g. [X].

Differential Expression Analysis (DEA). DEA was performed on the quality-controlled filtered [raw/normalized] counts using the limma (ver) [ref] workflow for fitting a linear model [formula] to identify features (genes/regions) that statistically significantly change with [comparisons] compared to the control group [reference levels] (intercept). Briefly, we determined normalization factors with edgeR::calcNormFactors (optional) using method [X], then applied voom (optional) to estimate the mean-variance relationship of the log-counts. We used blocking on (optional) variable [X] to account for repeated measurements, lmFit to fit the model to the data, and finally eBayes (optional) with the robust (and trend flag – optional for normalized data) flag to compute (moderated/ordinary) t-statistics. For each comparison we used topTable to extract feature-wise average expression, effect sizes (log2 fold change) and their statistical significance as adjusted p-values, determined using the Benjamini-Hochberg method. Furthermore, we calculated feature scores, for each feature in all comparisons, using the formula [score_formula] for downstream ranked enrichment analyses. Next, these results were filtered for relevant features based on the following criteria: statistical significance (adjusted p-value < [X]), effect size (absolute log2 fold change > [X]), and expression (average expression > [X]). Finally, we performed hierarchical clustering on the effect sizes (log2 fold changes) of the union of all relevant features and comparison groups.

Visualization. The filtered result statistics, i.e., number of relevant features split by positive (up) and negative (down) effect sizes, were visualized with stacked bar plots using ggplot (ver) [ref]. To visually summarize results of all performed comparisons, the effect size (log2 fold change) values of all relevant features in at least one comparison were plotted in a hierarchically clustered heatmap using pheatmap (ver) [ref]. Volcano plots were generated for each comparison using EnhancedVolcano (ver) [ref] with adjusted p-value threshold of [pCutoff] and log2 fold change threshold of [FCcutoff] as visual cut-offs for the y- and x-axis, respectively. Finally, quality control plots of the fitted mean-variance relationship and raw p-values of the features were generated.

The analysis and visualizations described were performed using a publicly available Snakemake (ver) [ref] workflow (ver) [10.5281/zenodo.7808516].

Features

The workflow performs the following steps that produce the outlined results:

• Differential Expression Analysis (DEA) steps:
  • (optional) calculation of normalization factors using edgeR::calcNormFactors.
  • (optional) calculation of precision weights to model the mean-variance relationship in order to make linear models "applicable" to count data (weighted least squares) using voom.
  • (optional) block on a "group" factor in case you have repeated measurements (generalized least squares).
    • example use-case: you have N donors and T timepoints for each donor and want to model donor specific information like age, but still want to account for the variable donor ie ~ timepoint + age + donor is overdetermined hence the formula becomes ~ timepoint + age and you "block" on donor.
  • fit linear models (ordinary least squares) with the design derived from the configured formula (expects "normal" data) using lmFit.
    • the fitted model object is saved (lmfit_object.rds) for alternative downstream analyses or manual inspection e.g., contrasts.
  • (optional) estimate variance "better" using eBayes, with the robustness flag (robust=TRUE), by looking across all genes (i.e. shrunk towards a common value) and compute moderated t-statistics.
    • (optional) eBayes with limma-trend (trend=TRUE)
  • extract all statistics for variables of interest (=configured comparisons) using topTable (eg coefficients/effect size, statistical significance,...).
  • save a feature list per comparison group and direction of change (up/down) for downstream analyses (eg enrichment analysis) (TXT).
    • (optional) annotated feature list with suffix "_annot" (TXT).
  • (optional) save feature score tables (with two columns: "feature" and "score") per comparison group using score_formula for downstream analyses (eg ranked enrichment analysis) (CSV).
    • (optional) annotated feature scores tables (with two columns: "feature_name" and "score") with suffix "_annot" (CSV).
• DEA result statistics: total number of statistically significant features and split by positive (up) and negative (down) change (CSV).
• DEA result filtering of features (eg genes) by
  • statistical significance (<= adjusted p-value: adj_pval)
  • effect size (>= absolute log 2 fold change: lfc)
  • average expression (>= ave_expr) in the data (to skip this filter use -Inf)
• Visualizations
  • filtered DEA result statistics ie number of features and direction (stacked bar plots)
  • volanco plots per comparison with effect size on the x-axis and raw p-value(rawp)/adjusted p-value (adjp) on the y-axis
    • highlighting features according to configured cut-offs for statistical significance (pCutoff) and effect size (FCcutoff)
    • (optional) highlighting features according to configured feature lists
  • hierarchically clustered heatmap of effect sizes (LFC) per comparison (features x comparisons) indicating statistical significance with a star '*'
    • using all relevant features (FILTERED)
    • (optional) using configured feature lists
    • in case of more than 100 features the row labels and significance indicators (*) are removed
    • in case of more than 50000 features no heatmap is generated
  • diagnostic quality control plots
    • (optional) voom mean-variance trend
    • (optional) intermediate mean-variance trend, in case of blocking and vooming
    • post-fitting mean-variance trend
    • raw p-value distributions (to check for p-value inflation in relation to average expression)

FYI

• Colons (":") in variable/group names (eg interactions) are replaced with double-underscores ("__") downstream.
• As of now we do not feature more complex contrast scenarios than are supported via topTable, but the fitted linear model is saved for downstream analyses.

Usage

Here are some tips for the usage of this workflow:

• limma usage and best practices are not explained. For deatiled documentation, tutorials and insctructions see Resources. To understand the implementation at hand see limma.R.
• Perform your first run(s) with loose filtering options/cut-offs and use the same for visualization to see if further filtering is even necessary or useful.
• Test minimal complex configurations on a subset of your data.
• Start with simple models and try to understand the results.

Configuration

Detailed specifications can be found here ./config/README.md

Examples

--- COMING SOON ---

Links

• GitHub Repository
• GitHub Page
• Zenodo Repository
• Snakemake Workflow Catalog Entry

Resources

• Recommended MR.PARETO modules for downstream analyses:
  • Enrichment Analysis for biomedical interpretation of results.
  • Genome Tracks for visualization of top hits.
• Bioconductor - limma includes a 150 page userguides
• R Manual on Model Formulae
• Bioconductor - RNAseq123 - Workflow
• limma workflow tutorial RNA-seq analysis is easy as 1-2-3 with limma, Glimma and edgeR
  • notebook
  • paper
• A guide to creating design matrices for gene expression experiments
  • notebook
  • paper
• voom: precision weights unlock linear model analysis tools for RNA-seq read counts
  • paper
• scRNA-seq DEA benchmark paper featuring limma-voom and limma-trend as valid/best methods
  • paper: Bias, robustness and scalability in single-cell differential expression analysis
  • code
• scRNA-seq DEA benchmark featuring limma-voom and limma-trend emphasizing pseudo-bulking
  • paper: Confronting false discoveries in single-cell differential expression
• alternative/complementary DEA method: Linear Mixed Models (LMM)
  • variancePartition
  • dream

Publications

The following publications successfully used this module for their analyses.

• ...

dea_limma's Issues

volcano plots legend has duplicate entries

config.yaml

volcano:
    pCutoff: 0.05
    FCcutoff: 0

consider capping heatmaps

• Get input from the team.
• cap heatmaps like in enrichment analysis e.g.,
  • simple cap ie LFC>cap_th -> set to cap_th
  • log? (log again the LFC?)
  • scale the heatmaps?

Consider: it is a breaking change as it requires a new config field...
same as in dea_seurat: epigen/dea_seurat#15

add example config, annotation and data

The examples should test all features and minimal features (e.g., without feature annotations)

logFC heatmap job does not complete if no DEGs detected

It would be nice to add some kind of a check and still label the job as completed, even if there are no DEGs (i.e. create an empty output file?)

DEA statistics barplot split of up and down by sign

To get a better overview of the number of differential features in each direction(!), up and down, the barplot could go literally up and down from zero. "down" features need a negative sign (-).

plot_df are the DEA statistics in long format with negative values for "down" values.

# plot config
direction_col <- list("up"="red", "down"="blue")

# plot
ggplot(plot_df, aes(x=groups, y=n_features, fill=direction)) + 
        geom_bar(stat="identity", position="identity") +
        scale_fill_manual(values=direction_col) +
        custom_theme +
        theme(legend.position = "none",
             axis.text.x = element_text(angle = 45, vjust = 1, hjust=1)
             )

consider including edgeR 4.0 workflow

https://www.biorxiv.org/content/10.1101/2024.01.21.576131v1

volcano plot: below indicate absolute LFC being used

below volcano plot replace
log2FC > 1
with
|log2FC| > 1

consider saving $E expression matrix that is used in the lmFit linear model

for downstream analysis or QC

support contrasts

• eg with contrast.matrix() & contrasts.fit() as strings in annot file https://wiki.bits.vib.be/index.php/Differential_expression_for_more_than_two_groups
• requires knowledge of the design matrix (columns) → 2 step process? first run, then check design matrix and specify in annot? what if more than one? annot only has one row…
• → does not seem feasible

consider determining the exact output files from the start for complete compatibility

DEA modules with explicit output to enable usage as module with subsequent modules (avoiding missing input exception) e.g., for enrichment analysis as input.
Requires loading all metadata files and explicitly defining final output files using limma /lmfit variable naming scheme.

Pro:

• enables smooth usage as a module with explicit outputs to be used for subsequent inputs

Con

• requires data dependent configuration/annotation -> considered bad practice/to be avoided
  • read up on why and how this translates to this use case

if done, do the same in dea_seurat.

configurable gene/feature lists to be plotted as LFC clustered heatmaps and highlighted volcanos (if present)

Snakemake v 7.21 requires output file in rule all

without adding a completion marker file ({}_succesfully_completed.txt snakemake would never start reporting "nothing to be done" on version 7.21