Skip to content

Prepare DESeq2 Specifications

Process Details

  • Name: Prepare DESeq2
  • Process UUID: J5fEA2uKhX2OEIsdNEsMEvPull9x1m
  • Process Group: Misc.

Overview

This process sets up the necessary inputs and options for running the DESeq2 analysis, including quality control, filtering, PCA plots, batch correction, DESeq2 analysis itself, significance thresholds, gene labeling, and more. It provides a comprehensive differential expression analysis workflow with extensive customization options for experimental design and visualization.

This process is implemented in Bash, which invokes a Python script (prepare_DESeq2.py) for orchestrating the DESeq2 analysis pipeline.

Key Functionality

  • Quality Control Analysis: Generates count distribution plots and reproducibility plots to assess data quality
  • Data Filtering and Normalization: Applies configurable filtering strategies and normalization methods with optional batch correction
  • Differential Expression Analysis: Performs DESeq2 analysis with customizable design formulas and statistical parameters
  • Visualization Generation: Creates PCA plots, volcano plots, MA plots, and heatmaps for result interpretation
  • Gene Annotation: Supports gene name conversion between different identifier systems (ENSEMBL, ENTREZID, SYMBOL)

Input/Output Specification

Inputs

Required Inputs

  • Count File

    • Description: Gene or transcript count matrix for differential expression analysis
    • Format: TSV

  • Run DESeq2

    • Description: Control parameter to execute the DESeq2 analysis workflow
    • Format: run_DESeq2

Optional Inputs

  • Group File

    • Description: Sample metadata file containing experimental group assignments and covariates
    • Format: TSV

  • Compare File

    • Description: Comparison specification file defining which groups to compare in differential expression analysis
    • Format: TSV

Outputs

  • Output File

    • Description: Processed results and analysis outputs from the DESeq2 workflow
    • Format: TSV

  • Postfix

    • Description: Identifier string for output file naming and organization
    • Format: postfix

Parameters & Settings

These parameters can be adjusted in the Foundry UI when running this process.

Quality Control Options

  • Create count distribution plot

    • Description: Generate plots showing the distribution of read counts across samples
    • Default value: true

  • Create reproducibility plots

    • Description: Generate all-to-all correlation plots to assess sample reproducibility
    • Default value: true

  • Create PCA plots

    • Description: Generate principal component analysis plots for sample clustering visualization
    • Default value: true

Feature and Filtering Options

  • Feature Type

    • Description: Type of genomic feature to analyze
    • Available options: gene (default), transcript

  • Filter Strategy

    • Description: Local filter considers only samples in the comparison. Global filter considers all samples within counts file. If Minimum Counts = 10 and an event has 5,5,11 counts respectfully in samples A,B,C it will be excluded in A vs B comparison with local filter, but included with global filter.
    • Available options: local (default), global

  • Minimal Counts

    • Description: Filter genes with less than this count. Used with min_samples parameter.
    • Default value: 10

  • Minimal Samples

    • Description: Filter genes which do not have at least min_count reads in this many samples. Used with min_counts parameter
    • Default value: 1

  • Minimal Total Counts per Sample

    • Description: Removes samples with less than this many total counts prior to DE. This is used to prevent very low quality samples from disrupting results.
    • Default value: 1000

  • Excluded Events

    • Description: Space, comma, or newline separated list of genes to remove prior to any processing. Separate reports will be created with/without exlusion of these events.
    • Default value: (empty)

Batch Correction Options

  • Run batch correction

    • Description: Perform batch correction prior to DESeq
    • Default value: false

  • Correction column

    • Description: Column used for batch correction.
    • Default value: batch

  • Biological variable column

    • Description: Column containing biological variable grouping.
    • Default value: group

  • Normalization Algorithm

    • Description: Normalization algorithm to use prior to batch correction.
    • Available options: MRN (default), TMM, RLE, upperquartile, none

PCA Plot Options

  • Data Transformation

    • Description: Transformation to apply prior to PCA (does not apply to data sent into DESeq). Default will use the rlog transformation for up to 50 samples and vst if there are more than 50 samples.
    • Available options: Default (default), rlog, vst, None

  • Color

    • Description: Column from metadata file by which to determine point color.
    • Default value: group

  • Shape

    • Description: Column from metadata file by which to determine the point shape.
    • Default value: (empty)

  • Fill

    • Description: Column from metadata file by which to determine point fill.
    • Default value: (empty)

  • Transparency

    • Description: Column from metadata file by which to determine point transparency.
    • Default value: (empty)

  • Label

    • Description: Column from metadata file by which to determine point label.
    • Default value: (empty)

DESeq2 Options

  • Run DESeq2

    • Description: Execute the DESeq2 differential expression analysis
    • Default value: true

  • Input Mode

    • Description: Samples to include when creating DESeqDataSet. 'All' will include all samples in the groups file. 'Comparison-only' will include only the samples from the respective comparison. DESeq2's dispersion calculated for each gene will be affected by which samples are included in the DESeqDataSet.
    • Available options: All (default), Comparison-only

  • Design formula

    • Description: DESeq2 design formula
    • Default value: ~group

  • Fit Type

    • Description: fitType used during DESeq2 analysis.
    • Available options: parametric (default), local, mean, glmGamPoi

  • Use batch corrected counts

    • Description: Use batch corrected counts in differential expression analysis
    • Default value: true

  • Apply shrinkage

    • Description: Apply shrinkage to fold change after DESeq2
    • Default value: true

  • Algorithm

    • Description: Shinkage algorithm to use.
    • Available options: apeglm (default), ashr, normal

  • Create volcano plots

    • Description: Generate volcano plots for visualization of differential expression results
    • Default value: true

  • Create MA plots

    • Description: Generate MA plots showing mean expression vs fold change
    • Default value: true

  • Create heatmap

    • Description: Generate heatmap visualization of differentially expressed genes
    • Default value: true

Significance Options

  • Adjusted p-value cutoff

    • Description: Adjusted p-value significance threshold required to be called differentially expressed.
    • Default value: 0.05

  • Fold change cutoff

    • Description: log2 Fold Change threshold required to be called differentially expressed.
    • Default value: 1

  • Adjusted p-value minimum

    • Description: Reassign adjusted p-values below this value to this value. Set at 0 to disable.
    • Default value: 0

  • Fold change maximum

    • Description: Reassign fold changes (log2) greater than this value to this value. Set to Inf to disable.
    • Default value: Inf

Labeling Options

  • Convert gene names

    • Description: Convert gene identifiers between different naming systems
    • Default value: false

  • From

    • Description: Name used in count file.
    • Available options: ENSEMBL (default), ENTREZID, SYMBOL

  • To

    • Description: Name to use in plots.
    • Available options: ENSEMBL (default), ENTREZID, SYMBOL

  • Number Labeled

    • Description: Number of significant genes to label. Set to Inf to label all. Set to 0 to turn labels off.
    • Default value: Inf

  • Highlighted Genes

    • Description: Space, comma, or newline separated list of genes to label regardless of significance.
    • Default value: (empty)

  • Create additional volcano plots for highlighted genes only

    • Description: Generate separate volcano plots focusing only on highlighted genes
    • Default value: false

  • Create additional MA plots for highlighted genes only

    • Description: Generate separate MA plots focusing only on highlighted genes
    • Default value: false

References & Resources

  • Tool Documentation: Contact the team for details on prepare_DESeq2.py
  • Related Papers: Love, M.I., Huber, W., Anders, S. (2014) Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology, 15, 550. https://doi.org/10.1186/s13059-014-0550-8