Hi,
I am working on differentially expressed genes analysis of RNA-seq data extracted from multiple but related experiments:
experiment
DESH21_E
DESH40_S
DESN21_E
DESN21_S
DESN40_E
DESN40_S
ES21_E
ES21_S
ES40_E
ES40_S
where, DES and ES are different in vitro cells; 21/40 are concentration (low/high), and H/N are additional stress (high/normal). All experiments have three replicates.
I want to use DESeq2 and compare several experiments, for example:
- DESN21_E vs DESH21_E
- DESN21_E vs DESN21_S
- DESN40_S vs DESN40_E
- DESN40_S vs DESH40_S
However, DESeq2 takes only the first level in the metadata to perform the comparisons.
An example:
dds <- DESeqDataSetFromMatrix(countData = counts,
colData = meta,
design = ~experiment)
dds <- DESeq(dds)
resultsNames(dds)
[1] "Intercept" "experiment_DEH21_vs_DESH21_S" "experiment_DEH40_vs_DESH21_S"
[4] "experiment_DEN21_vs_DESH21_S" "experiment_DEN40_vs_DESH21_S" "experiment_DESH21_E_vs_DESH21_S"
[7] "experiment_DESH40_S_vs_DESH21_S" "experiment_DESN21_E_vs_DESH21_S" "experiment_DESN21_S_vs_DESH21_S"
[10] "experiment_DESN40_E_vs_DESH21_S" "experiment_DESN40_S_vs_DESH21_S" "experiment_E21_vs_DESH21_S"
[13] "experiment_E40_vs_DESH21_S" "experiment_ES21_E_vs_DESH21_S" "experiment_ES21_S_vs_DESH21_S"
[16] "experiment_ES40_E_vs_DESH21_S" "experiment_ES40_S_vs_DESH21_S" "experiment_S21_vs_DESH21_S"
[19] "experiment_S40_vs_DESH21_S"
How can I use the whole count matrix and perform the different comparisons without re-run DESeq changing the levels in the metadata to carry every comparison?
Thanks in advance,
Juan
Hi,
If you only want to get pairwise comparisons, then you can use
contrasts
as described in theDESeq2
manual: https://www.bioconductor.org/packages/devel/bioc/vignettes/DESeq2/inst/doc/DESeq2.html#contrastsAntónio