Talk:Volcano plot (statistics)

Terminology

What is a "fold-change"? The usage here seems different from that given in fold change, which seems to indicate a value that is always greater than one, while here the example shows that values can be less than one. Perhaps something more direct needs to be said here anyway. Melcombe (talk) 12:19, 4 January 2010 (UTC)

the usage of fold change is the same, but one plots the log of the fold change on the x-axis. the log is used so that changes in both directions (up and down) appear equidistant from the center (0). so if you're using log2 on the x axis, a point that has a fold change of 4 will appear at +2 and a fold change of 0.25 will appear at -2, which is what you want because these fold changes are identical, just in opposite directions. does that make sense? i added a sentence to the article to this effect, but let me know if you think it needs further clarification. Roadnottaken (talk) 15:48, 4 January 2010 (UTC)

But fold change says "It is the ratio of the final value and initial value (B/A), if the final value is larger." My interpretation of this is that you get either B/A or A/B depending which is larger. That would certainly be a possible definition, but obviously it is not what is wanted here. I have not come across "fold change" as a specific terminology before, so I don't know what is "usually" meant. The fold change article is rather new and concerns were expressed about it in its edit history, so perhaps that article justs needs to be made more meaningful. Melcombe (talk) 11:21, 5 January 2010 (UTC)

Yah, my impression is that "fold-change" is a pretty generic term like "difference" or "ratio", but if statisticians have a precise definition of it and I'm using it incorrectly, I'll be happy to change its use in this article. Roadnottaken (talk) 18:47, 5 January 2010 (UTC)

The term "fold change" isn't used in mainstream statistics, but I see by an internet search that it is in genomics. The meaning, as used in this article is clear enough ... the problem is with the fold change article. Melcombe (talk) 11:39, 8 January 2010 (UTC)

p values are normally distributed?

I just made an edit. A previous editor said the p-values are normally distributed, but this is unclear. For ex, iinm, with random data, p-values assumea uniform distribution across the (0,1) interval. I believe the statement was meant to say something about the distribution of the p-values against the log-fold changes, but this wasn't explained legibly, and I am not enough of a statistician to explain it better. Parabolas do sometimes appear in volcano plots (seen here http://resources.qiagenbioinformatics.com/manuals/clcgenomicsworkbench/current/index.php?manual=Volcano_plots.html ) , but someone else will have to explain why that is. Flies 1 (talk) 18:34, 15 January 2021 (UTC)

Source for the interested

This source may be useful to describe such plots. Jo-Jo Eumerus (talk) 14:45, 17 November 2023 (UTC)

With respect, the source quoted (an American Chemical Society publication) is not useful for finding out about the meaning of volcano pots in the modern "omics" research fields. While the ACS source is interesting in itself for chemists who study catalytic behaviour of metal surfaces (as in catalytic converters), the plots discussed in that article look different (the mountainous aspect of volcanoes) from those in the Volcano-plot Wikipedia article (what shoots out of an active volcano), resulting in the visual effects being related by one looking like the other, but upside-down. Moreover, the catalysis oriented chemical usage of these volcano plots do not intend to look after the statistical interest of the Wikipedia article, which balances probabilities against the magnitude of the deviation from the norm, that is intended to be shown in the p vs. fold-change volcano plots of the article. In short, they are a different kind of plot that unfortunately got the same name. — Preceding unsigned comment added by 2A01:E0A:149:BEB0:ACED:45BE:E2C9:67E (talk) 10:49, 13 November 2024 (UTC)