Hello Simone,

There are several methods that can be used. But we need more information on your dataset to properly answer the question. E.g. number of species, number of columns (i.e. temperature value), whether only the shape of the curve is important or also the other characteristics of it, such as heigh (abundance), maximum point of the Gaussian curve, inflection point of the sigmoid curve etc.

If you have a lot of species and you can preliminarily categorize a subset of them with expert knowledge, then I would use an artificial intelligence method, such as neural network, on the raw dataset.

If you have fewer species or no preliminary categorization, then I would calculate some important characteristics of the temperature-abundance relationship for all the species, and use these calculated variables in a clustering or ordination method. E.g. you can calculate the slope of temperature-abundance curve (for different temperature intervals) by fitting a simple linear model, or calculate the skewness, the maximum point etc.

HTH,

Ákos

Ákos Bede-Fazekas thank you for replying to my question, well for me it would be interesting just the shape of the curve, it doesn't matter the rest, so the goal is clustering bacterial taxa with a similar curve (e.g. logistic, sigmoidal, quadratic plateau, exponential and so on). The database is made by about 300 lines (species) and the columns are 10 (temperature levels)

Dear Simone Anzà ,

Well, 300 species are not enough for a supervised neural network or other artificial intelligence algorithm. So among the two methods I mentioned in my previous post only the second one is feasible. If the hight of the curve is not important, then I would standardize/Studentize all rows, i.e. for all the values I would subtract the mean of the 10 temperature values and then divide it by the standard deviation of them.

Have a nice week,

Ákos

Check https://courses.lumenlearning.com/boundless-microbiology/chapter/temperature-and-microbial-growth/

Also check https://www.britannica.com/science/bacteria/Classification-by-morphology-biochemistry-and-other-features

Hello Anza, you can also use this

https://www.slideshare.net/viveksharma1174/classification-of-bacteria-based-on-temperature-ph-and-oxygen-reqiurementdoc-pdf

thank Kiprotich Kiptum and Chinaza Godswill Awuchi for your answers but what I am looking for is a stat method that I can use to cluster bacterial according to specific functions