I notice in the latest DESeq version that it can handle GLM type experiment designs now, but the example still only does a pairwise comparison. Is it possible to do a time series using:
?
I got that from the DESeq reference guide but the guide doesn't actually specify what 'count ~ 1' does. I assumed it either uses condition 1 as the base or treats each sample separately instead of grouping into conditions, which does it actually do?. Is my use of it here correct?
Code:
fit1 = fitNbinomGLMs( cdsFull, count ~ condition ) # in my case times: 0,6,12,24,48,72 hours fit0 = fitNbinomGLMs( cdsFull, count ~ 1 ) pvalsGLM = nbinomGLMTest( fit1, fit0 ) padjGLM = p.adjust( pvalsGLM, method="BH" )
I got that from the DESeq reference guide but the guide doesn't actually specify what 'count ~ 1' does. I assumed it either uses condition 1 as the base or treats each sample separately instead of grouping into conditions, which does it actually do?. Is my use of it here correct?