PossionDis method for identification of differentially expressed genes.

Question

[M]

Hello,

I recently found a paper where the authors claim they used a PossionDis method for identifying differentially expressed genes. This made me curious, as I never heard about it before. When I google PossionDis, I found only papers that have a poorly description of the analysis (one sentence) that seems to be copy pasted from one paper to the other, no description of software used or anything. All of them are referenced to an article from 1997, where PossionDis is not even mentioned, which makes me suspect that somebody at some point misspelled Poisson, and that error was propagated through the papers.

Have any of you heard about it or have any thoughts on it? Where can one even find an implementation of this tool?

[Bastien Hervé]

Just by curiosity, which paper did you first find with this PossionDis mention ?

[M]

it was this one

Answer 1

[dsull]

Never heard of PossionDis; maybe it's a custom tool that someone wrote (based on that 1997 method) that happens to be shared among a few labs.

I would not bother looking any further than that. Just run your analyses using a method like DESeq2 or edgeR like everyone else; those methods have actually been thoroughly researched, tested, peer-reviewed, and benchmarked, and account for things that older methods (e.g. the 1997 method) do not.

As for the 1997 article you cite, I did find this paper: https://academic.oup.com/bioinformatics/article/35/1/170/5055407 (from the original authors) which presents a tool that implements that method. It's not the "PoissonDis" tool but it's an implementation of the 1997 method.

Answer 2

[Bastien Hervé]

I am googling for like 20 minutes now and this is super interesting.

It looks like only a few labs had mentioned PossionDis recently. I haven't done a thorough lab network but it looks like most of them are coming from China with the same constellation of researcher cited. I guess your were right saying this is making reference to Poisson distribution as the authors always mentioned "Possion distribution".

Of note, one can also find some sporadic misspelling of Poisson distribution already in 1975 : https://pubmed.ncbi.nlm.nih.gov/?term=Possion&filter=years.1975-1977

Also on pubmed there is an erratum here for one of the paper : https://pubmed.ncbi.nlm.nih.gov/36308862/

Really interesting finding, maybe it would be advisable to contact the authors and journals involved.