Welcome to the New Seqanswers!

Welcome to the new Seqanswers! We'd love your feedback, please post any you have to this topic: New Seqanswers Feedback.
See more
See less

Getting normalized counts

  • Filter
  • Time
  • Show
Clear All
new posts

  • Getting normalized counts


    I've tried to find differential expression with edgeR and DESeq2. These packages normalize my count data first, but they do not automatically provide normalized read counts to the end user. Any ideas how to obtain them through a series of R commands? Where could I look for this?

    Thank you in advance!

  • #2
    For DESeq2, use the counts function, passing as arguments the DESeqDataSet object created by your call to the function DESeq, and the argument, normalized=True.

    The function will return an integer matrix.
    In the following example, I convert the matrix to a data frame.

    # Create a DESeqDataSet object, and perform calculattions.
    dds <- DESeqDataSetFromHTSeqCount(sampleTable=sampleTable, directory="../../htseqcount", design= ~ condition)
    dds <- DESeq(dds)
    # Calculate and save the annotated normalized counts.
    normalized.counts <- dds, normalized=TRUE ))
    As to where you can get this information, you should read the DESeq2 vignette.
    It is quite complete, and updated regularly.
    After reading it a dozen times , I have a clearer idea how to use DESeq2, especially for more complex analyses, like multi-factorial analyses.
    Last edited by blancha; 05-31-2014, 08:38 AM.


    • #3
      For the edgeR verions, just:
      normalized.counts <- 1e6 * cpm(dge)
      where "dge" is a DGEList.