Using vegan to calculate alpha diversity metrics within the tidyverse in R (CC196)

[@AlertedDonkey42]

Literally just started working with some invertebrate community data today and saw this. It's helpful! Thanks!

[@Riffomonas]

Wonderful!

[@SalmaSarker-i1p]

Dear Patrick, I hope this email finds you well. I am currently working on soil microbiology and analyzing data in R. Your videos have been incredibly helpful for me as a novice researcher. I have been closely following your Raffomonas channel and your publications on the importance of rarefaction for ecological microbial data, which have been immensely insightful. Based on your recommendations, I implemented empirical rarefaction, performing 1,000 iterations with my desired sequencing depth. From this, I derived the mean richness and standard deviation values for all my samples. However, I am now looking to move forward with beta diversity analysis and calculating various indices of alpha diversity. This requires an ASV table, but the empirical rarefied data I currently have does not include an ASV column. I managed to combine my treatment metadata with the empirical data, and I also have a phyloseq object from before the rarefaction process. I was wondering if it is possible to perform empirical rarefaction directly on my phyloseq object. If so, what output should I expect? Would it provide an ASV table along with mean richness and other metrics? Additionally, I have read in your paper that “Rarefaction repeats the subsampling a large number of times (e.g., 100 or 1,000 times) and calculates the mean of the alpha or beta diversity metric over those subsamplings; rarefaction is implemented with the vegan R package using the rarefy or avgdist functions.” Could you please provide more guidance on how I should proceed with alpha and beta diversity analysis using empirical rarefied data? Your insights would be invaluable for my project, and I sincerely appreciate your time and expertise.

[@Riffomonas]

Hi Salma - I really know nothing about phyloseq :) I'd encourage you to use mothur instead. You can see how we use rarefaction for alpha and beta diversity using OTUs, phylotypes, and ASVs in the MiSeq SOP - mothur.org/wiki/miseq_sop/

[@theresadavenport8346]

Theis video series is so helpful! I am following along with my own data, highly recommended!

[@ggimdone4138]

Stunning. Learnt alot from this vid. Thanks man, super clear!!!

[@Riffomonas]

My pleasure - thanks for watching!

[@kishormaharjan5313]

Hi, this is great. I'm struggling with creating the function for Pielou’s Evenness Index and including that one along with richness, Shannon, Simpson. Please help.

[@abdoulieb.jallow996]

thank so much for the video. what about if sample autumn, winter and spring in a river at 3 different locations and you were to look at this by locations and season in the year?

[@Riffomonas]

I'd encourage you to check out the summary.single function in the mothur package to rarefy all samples to the same level of sequencing effort

[@allmywhat]

Great content! Could you show a video on how to incorporate the rarefied data from your previous videos into plotting alpha diversity to compare it to non-rarefied data?

[@Riffomonas]

Thanks Michael - I do this a few times over the series on rarefaction. Non rarefied data are more variable, larger and produce spurious significant differences more often when effects are confounded with sampling effort

[@hope2024-g]

Thank you so much Sir for your awesome video! Could you show us Generalized additive model using mgcv package with interpretation using a data contain both continues and categorical variables plz?

[@Riffomonas]

Thanks for watching! Unfortunately that’s not a tool I use so it’ll probably be awhile before I make a video with that

[@svetlinavasileva109]

Hi Pat, thanks so much for your videos, they are pure gold! I was wondering how can I apply this to data in another format. I have a dataframe (df) with samples as rows and taxa as columns. I tried this code, but it doesn't work. df%>% summarize(richness=richness(x), shannon=shannon(x), simpson=simpson(x), invsimpson=1/simpson(x), n=sum(x)) I guess, I am not sure what to put instead of (value) as in your code. Or would I need to transform the data?

[@Riffomonas]

Thanks for watching! You’d still need to pivot longer to get columns of samples, taxa, and counts. Then group by samples and then the summarize function you have

[@svetlinavasileva109]

@@Riffomonas Thank you!

[@KarizmaFahlevy]

Hi, thank you very much for the videos !! Im still new in R, and happy to learn Do you have any lecture video for functional diversity, traitspace, CWM using the FD packages?

[@Riffomonas]

sorry but i don't

[@KarizmaFahlevy]

@@Riffomonas Thanks

[@emekaemmanuel5258]

Does the Vegan function calculate Observed ASVs and Faith's Phylogenetic Diversity (PD) index?

[@Riffomonas]

Not that I know of. I calculated PD using mothur.

[@deodatusruganuza5254]

I am working on human GI nematodes which has one species, that has multiple ASvs are diversity measure important??

[@AntaraKulkarni]

This is a great video, thank you so much! I was wondering about the y axis in the calculation of a Shannon Diversity. If n represents the total abundance of all the sites, wouldn't it keep increasing for all metrics?

[@josenicolasperez-garcia8119]

Can yo help me, in Anosim test how I can have the pairwase R and p values? I have the general result in Rstudio but not the pairwase result

[@Riffomonas]

You might need to assign the output to a variable and then run str on that variable

[@efratsharon1294]

Thank you, It would be helpful to add statistics on top of the alpha diversity metrics calculated. Thanks

[@Riffomonas]

I’m not 100% what you mean but you can add any text to a plot by adding a geom_text to the existing plot.

[@efratsharon1294]

@@Riffomonas I meant calculate the significance of comparing between Alpha diversity of healthy and sick (or other conditions)

[@heidiyeh7585]

@@efratsharon1294 I second this, specifically how to run an ANOVA that shows whether the groups you were analyzing for the first 10 minutes of this video have a statistically significant difference in their alpha diversity values

[@mohamedmdaini8143]

Thank you for the amazing tuto, but the "%>%" doesnt really work in my case, it always shows "could not find function "%>%", any solutions? Thanks

[@tommajor3725]

you need to install the dplyr package install.packages("dplyr") library(dplyr)

[@Riffomonas]

be sure that you've run library(tidyverse) or library(magrittr) to get the %>% to work. Alternatively, you could use the base R pipe - |>

[@mohamedmdaini8143]

@@Riffomonas thank you

[@cristianjaviermena6955]

Thanks you Patch!

[@bjmoreno100]

what is the structure and format of the data and tables that you use to perform this analysis ?

[@bjmoreno100]

i have a table of tha that contains a information of each individual in one row , and i only want to select group, name and value when the result of value is the count of name times apear in one group.

[@Riffomonas]

The rows are samples and the columns are taxa. You can see the file if you go to the link in the description and then to the GitHub link

[@spencersiddons5577]

Hi there. How can I import the dataset from github into R? I select "Download" in the github file and it brings up the data in a table (the .shared file) in a new webpage, but I cannot figure out where to go from there. Thank you!

[@Riffomonas]

Thanks for watching Spencer. If you follow the video at riffomonas.org/code_club/2022-01-24-catching-up you'll see how to get the data you need.

[@spencersiddons5577]

@@Riffomonas Thank you!

[@upheaveworker2108]

So basically, the alpha diversity is in our study plot, but in beta diversity, it must be compared to alpha diversity in other region/continent/island/country, mustn't it?

[@Riffomonas]

beta takes into account who is there and in what abundance. alpha doesn't care. two communities could have the same number of taxa (alpha), but completely different taxa (beta)

[@meseretmuche6984]

remarkable lecture Q: how can i calculate the hill number diversity index in r (q=0, q=1, and q=2)

[@Riffomonas]

Thanks for watching - i don't know that vegan does hill numbers (i've never used them in my work). You could certainly write your own function to do it in R

[@jotaro.10]

i am a beginner, how to get that "x" in our data?

[@Riffomonas]

Sorry, I'm not sure where you are referring to - can you point me to where in the video you are asking about?

[@jotaro.10]

Like the one in richness

[@Riffomonas]

@@jotaro.10 x is a "dummy" variable that is used to calculate the alpha diversity metric. If you look at around 4:00 you'll see the richness function is called with the values from the "value" column in the dataframe

[@SeguC48]

Beautiful !!!

[@Riffomonas]

Thank you! Glad you found the channel 🤓

[@hermanediesse6035]

the best video ever