Changes between Version 63 and Version 64 of SOPs/rna-seq-diff-expressions

Timestamp:

11/03/22 09:53:38 (2 years ago)

Author:

gbell

Comment:

--

SOPs/rna-seq-diff-expressions

@@ -118 +118 @@
 
   * **Gene filtering**
-    * Remove from the analysis any genes with 0 counts across all samples.  Some analysis tools do this themselves.
-    * Remove counts for any genes we want to classify as contaminants or simply "not interesting" such as ribosomal RNAs (if these are included in the GTF gene annotation file).
+    * One used to remove from the analysis any genes with 0 counts across all samples.  Now most analysis tools do this themselves, so we skip this step.
+    * Remove counts for any genes we want to classify as contaminants or simply "not interesting" such as ribosomal RNAs or pseudogenes (if these are included in the GTF gene annotation file).  If in doubt, leave them in.
 
   * **Normalization**
@@ -132 +132 @@
        * Effective library size (a more complex, but probably more valid method included in programs such as edgeR and DESeq)
      * '''Differential expression statistics packages can output a matrix of normalized counts (typically using the method recommended for the accompanying statistics), so typically no additional normalization needs to be done''' (unless one want to perform further normalization, such as using gene length).
-     * As with microarray normalization, be aware of the assumptions of each method and choose the method(s) which are most valid with your experiment.
+     * As with all types of normalization, be aware of the assumptions of each method and choose the method(s) which are most valid with your experiment.
      * Origin of recommendation for upper-quartile normalization: [[http://www.ncbi.nlm.nih.gov/pubmed/20167110|Bullard et al., 2010]]
      * Use DESeq, DESeq2, or see [[http://jura.wi.mit.edu/bio/scripts/R/normalize_DGE_matrix.R|normalize_DGE_matrix.R]] for a command-line script for count normalization