out of 27464 with nonzero total read count
adjusted p-value < 0.1
LFC > 0 (up) : 0, 0%
LFC < 0 (down) : 0, 0%
outliers [1] : 0, 0%
low counts [2] : 0, 0%
(mean count < 0)
[1] see 'cooksCutoff' argument of ?results
[2] see 'independentFiltering' argument of ?results
•
written
by
sushumitha6 •
2 answers
There is a tradeoff. When we try to be precise, the explanations become a little unwieldy and make the definition hard to remember. When we make the explanation simple, the wording is ambiguous.
Here is a definition that I like, tries to reconcile the two extremes:
The p-value is the probability that the observed difference between two groups is due to chance.
I find this just as succinct and less easily misinterpreted: The p-value is the probability that random chance could produce the observed difference
the phrase "due to chance" has no connection to the magnitude of the p-values at all. Observing smaller or larger probability does not affect the interpretation of the second part of the sentence.
"due to chance" is just a succinct way to say "caused by random chance while the null hypothesis is true".
I like Jeremy Leipzig formulation as well:
The p-value is the probability that random chance could produce the observed difference
We have here a formulation that, while incomplete, captures the salient points. On the other hand, we could also use seemingly precise, words that, in turn, need to be defined separately, and remain hard to recall and interpret correctly. In my opinion, using words like null hypothesis is true is the primary reason that most people don't know what p-values are even at a simpler level.
Let me also posts the Biostar Handbook's theorem on p-values:
Nobody understands p-values. We only differ in the degree of our misinterpretation.
Log in to answer this question.
-
Mean count in DESeq
written by prerna •
-
DESeq2 results vs. shrinkage results
written by hagl •
-
DESEq2 results : low counts too many genes
written by shome •
-
Decrease the low counts
written by rufusek500 •
-
lfcShrink MA plot comparison
written by user230613
-
How to select differential expressed genes using DeSeq2?
written by Aynur •
-
How to generate a dataframe of deseq2 result excluding 'low counts' and 'outliers' ?
written by tianshenbio
-
model design in rna seq - deseq2
written by acebolladaso.iacs •
-
number of genes for co-expression network
written by Edalat •
-
Mean count in DESeq2 result summary
written by aw17 •
https://lmgtfy.app/?q=p-value&iie=1
Jorge, I don't think we should be using lmgtfy on the forum, we can ask people to simply Google stuff. lmgtfy is kind of passive aggressive.
I'm deleting your comment so this conversation is hidden for now. We'll undelete it once we decide on something.
Fair enough, Ram. I was just surprised how anyone could think that the best way to get his first p-value knowledge would be to post a question in a bioinformatics forum rather than performing a simple google search to read the first 2-3 results. Anyway, I'm also surprised how such basic question have arisen a short yet fruitful discussion, so I'll close my mouth right now and I'll keep on learning.
I agree with both your points but saying "Google search should give you a bunch of articles" is better and definitely more professional than lmgtfy. lmgtfy turns people off from seeking help.
Hi sushumitha6, it is generally expected that you read available material first, and for a tool like DESeq2 that would be the paper and the vignette at Bioconductor which will contain all the information necessary to understand this output.
For general starters I strongly recommend the StatQuest series over at YouTube:
As well as discussing the attempts you have made to investigate this yourself before post here (as discussed by @ATPoint below). Can you please elaborate some of the context for your question? Why is the experiment, why are you conducting this analysis? Otherwise one cannot interpret the result. Further, a question like yours, without context, could be mistaken for someone asking for help with a homework or exam question.