plotCutoff parameter in fcPlot

[elisabettasciacca]

hi Kat,

I don't know if it's me, but the role of the plotCutoff option in the fcPlot function doesn't look to be very clear . I think it would be good if you can explain better what it is meant for. I was trying to remove all the dots that are below my significance threshold (q < 0.05), I thought this option was able to help me but maybe it's not the case.

[KatrionaGoldmann]

Hi Elisabetta,

I've tried to clear this up in the latest release. The cutoff acts across all p-values (standard as well as interaction) so genes are excluded if all terms are considered insignificant. In order to exclude genes by a specific p or q-value you can intercept the glmmSeq object, for example:

# subset to only genes where p_timepoint <= 0.01
results@countdata = results@countdata[results@stats$P_Timepoint <= 0.01, ]
results@predict = results@predict[results@stats$P_Timepoint <= 0.01, ]
results@stats = results@stats[results@stats$P_Timepoint <= 0.01, ]

fcPlot(glmmResult=results,
       x1Label="Timepoint",
       x2Label="EULAR_6m",
       x2Values=c("Good responder", "Non responder"),
       pCutoff=0.1,
       labels=labels,
       useAdjusted = FALSE,
       plotCutoff = 0.1)