Видео

I'll need to look at what I load more closely this weekend, but you may need ggpubr loaded. Sorry for the lack of clarity in the video!

@@ecomattp for the stat_compare_means you need the ggpubr package ^^

Hello! To get you an answer sooner than later check this video out: ruclips.net/video/Uyof3S1gx3M/видео.html they do a good job explaining this exact item of interest. I hope this helps!

@@ecomattp Thank you. This is very helpful.

Thanks! Appreciate the praise, I have a tutorial on gganimate coming up that I'd be interested to hear your thoughts on/any feedback is welcome 😊 cheers!

I will work to get this on my blog this weekend!

That's a good question, in this regard it would be comparing accessions (ignoring treatment). If you wanted to isolate treatments in their own figures you could create the same comparison list and it would compare accessions within the NF treatment or the WF treatment (or whatever you happen to name your treatments). If you want to try and keep everything in one panel of figures I would suggest looking at faceting your figure.

Hi! After giving it some thought on the best way to share my code I made a quarto blog and the code for this video is the first to be put on it! you should be able to find it here: rpubs.com/ecomattp/1137343

Hi! After giving it some thought on the best way to share my code I made a quarto blog and the code for this video is the first to be put on it! you should be able to find it here: rpubs.com/ecomattp/1137343

Hello! To answer your second question, yes you can. In the line stat_compare_means you would need to add annotations=*** to your specific list of what you want to compare. This link should provide further detail! stackoverflow.com/questions/49008460/add-p-value-or-asterisk-to-ggplot As far as using a bootstrap method, I have not done that myself, but that's not to say it is not possible. Here is a pdf of a document I found that may be useful: pages.stat.wisc.edu/~larget/stat302/chap3.pdf I hope this helps!

It would be comparing only Afghan and Viridis, if you wanted to compare across all accessions I would look at using an ANOVA but also consider distinguishing between treatments.

Sorry for the late response! Good question. If you wanted to compare everything I would recommend an ANOVA (analysis of variance) which will compare everything. However, I just did control to each different treatment in this video to demonstrate customizable comparisons. Therefore, if you wanted to make other types of comparisons like I did in this video, the process is the same as shown, you would just need to create the list of comparisons you want to display.

Hey, sorry about that! Not sure why the video is blurry. I re-recorded the video here: ruclips.net/video/kfvvRUXAqrI/видео.html this should be much better!

I apologize for such a delay in response. I hope you have solved your issue! But if not here is a link to a possible solution: community.rstudio.com/t/quarto-render-button/143423

Thanks, Matthew.@@ecomattp

Hey, sorry about that! Not sure why the video is blurry. I re-recorded the video here: ruclips.net/video/kfvvRUXAqrI/видео.html this should be much better!

Sorry for only getting to this comment now! Check out my other video, hopefully it helps ruclips.net/video/IKoh4oiKtsg/видео.html

This could be anything from network connection to needing another package installed. Without seeing the code I cannot fully assess what the issue might be. If you copy and paste the error into google that may help.

Hey Stefano, Great question! You can create the diagram without doing the binary transformation, you can simply skip that line of the code and the rest should still run smoothly!

​@@ecomattp Hey Michael, I tried skipping the binary transformation line but now instead of having 3 circles (3 treatments) it only creates one. Since I want to compare the total abundance of taxa between treatments, I am also skipping two other steps that you included: 'remove taxa that sum up to 0' and 'remove taxa that sum up to less than 4 for each rep'. Could that be an issue?

Thank you! I appreciate hearing that!

If you highlight the section of code you want to run and press control + enter it should appear to show the progress of running your code. If there are errors it would become red. Good question!

You can use the same exact syntax, just only tell it to select the single set of data you want to show in the line plot. For instance in this video I tell it to select both 2018 AND 2019 in the geom_line section. If I remove the geom_line for 2019 it would show the timeline with just 2018 data. You will also need to remove the geom_point line that calls on the 2019 data or else the data will show up still without a line. I hope this helps!

Hi, thanks for pointing this out, not sure why the beginning is like this, but around 1:15 it improves on my end! Hopefully that is the case for you as well.

Hello! You should be able to add it into stat_compare_means. It might be that you add it in as a t.test rather than an ANOVA. It would look something like this: stat_compare_means(method="t.test"). I hope this helps. If not here is a link to break down the components of adding comparisons in ggplot: rpkgs.datanovia.com/ggpubr/reference/stat_compare_means.html

Very nice video, how can I do for two factors anova?

@@rajendraregmi2892 Great question. You would need to run ggplot through a series of pipes "%>%" leading with your data frame, in this case DN_Master_Data %>% (insert the ggplot script to create your boxplot). Then after you create your plot you can do a summary function of "anova" as such anova(lm(pH ~ Treatment, DN_Master_Data)). This creates a linear model and you can further detail interactions within this function. To avoid being too long winded, here's a link to a detailed explanation of how you can achieve this! I hope this helps: publish.illinois.edu/liangf/teaching/stat-424/rcode-two-way-anova/

@@ecomattp Thank you. I will go through this link and let you know if I need more help. Thanks

The [,-1] represents the removal of data with the parameters we will set in the remainder of this line. In this case we are saying "if the data across all of our samples sums up to 0, remove it" hence why 0 is listed after [,-1]>0. I hope this helps

Is this data set available so that we can follow along?

@@nf3922 Warnings are different from error messages in RStudio, so ggplot2 should still work! The dataset is not available unfortunately, but it should work similarly with pre-loaded datasets such as iris in RStudio. Keeping that in mind I will aim to use pre-loaded data in future videos! I hope this helps.

Here is the script including the packages I typically use for all of my figures: install.packages("dplyr") # one of the core packages of the tidyverse, extremely helpful for dataframe manipulation install.packages("tidyr") # used to help reshape data into tidy data that is easier to manipulate install.packages("permute") # allows you to perform permutations install.packages("lattice") # aids in graphics to provide better defaults to easily display multivariate relationships install.packages("vegan") # this is your go-to community ecology package. Used for ordination, diversity analysis and more. install.packages("goeveg") # aids your vegan package in ordination diagrams, species response curves, and rank-abundance curves install.packages("ggplot2") # your optimal package for creating excellent graphs and figures install.packages("tidyverse") # a collecton of R packages designated for data science (the parent package to ggplot2, dplyr, tidyr and more) install.packages("nlme") # for Gaussian linear and nonlinear mixed-effects models install.packages("ordinal") # facilitates analysis of ordinal data install.packages("ggsignif", type="win.binary") # significance brackets for ggplot2 install.packages("ggpubr", type="win.binary") #This is the package for running anova with the stat_compare_means line install.packages("multcompView") install.packages("ggthemes") #now that we have installed our packages we need to access them by loading them in our library. library(ggpubr) library(ggplot2) library(dplyr) library(tidyr) library(permute) library(lattice) library(vegan) library(goeveg) library(tidyverse) library(nlme) library(ordinal) library(multcompView) library(ggthemes) #making a timeline with multiple lines and a legend precip <- read.csv("INSERT YOUR DIRECT PATH TO CSV FILE HERE MAKE SURE YOU USE FORWARD SLASHES") #organize your data by month because the default is alphabetical precip$Month <- factor(precip$Month, levels = c("Jan.", "Feb.", "Mar.", "Apr.", "May", "June", "July", "Aug.", "Sept.", "Oct.", "Nov.", "Dec.")) ggplot(data = precip, aes(x=Month, group=1))+ geom_line(aes(y=Twenty18, color ="2018"), size=1)+ geom_line(aes(y=Twenty19, color ="2019"), size=1)+ scale_color_manual(values = c("2018"="darkred","2019"="steelblue"))+ geom_point(aes(y=Twenty18), size=2)+ geom_point(aes(y=Twenty19), size=2)+ theme_classic()+ theme(axis.title.x = element_blank())+ labs(color = "Year")+ ylab("Precipitation (mm)") I hope this helps!

These are all of the standard packages I use: (ggpubr is crucial for this method) library(ggpubr) library(ggplot2) library(dplyr) library(tidyr) library(permute) library(lattice) library(vegan) library(goeveg) library(tidyverse) library(nlme) library(ordinal) library(multcompView) library(ggthemes) check to see if you have these installed! If that doesn't work, here's a website that breaks down the same process: www.sthda.com/english/articles/24-ggpubr-publication-ready-plots/76-add-p-values-and-significance-levels-to-ggplots/

These are all of the standard packages I use for any project I start, not all will directly relate to the timeline you are trying to make, but in general I have found helpful. library(ggpubr) library(ggplot2) library(dplyr) library(tidyr) library(permute) library(lattice) library(vegan) library(goeveg) library(tidyverse) library(nlme) library(ordinal) library(multcompView) library(ggthemes) I hope this helps! If you continue to have issues, let me know!

@@ecomattp Thanks a million Matthew. You are such a great person. the date in x axes works now, but still a problem existed. when I use geom_line, the points are not linked together. every two neighbor points are linked vertically. If you know how could I overcome this problem, I would really appreciate your help. I really got headache

@@OMAR-ce1bj I'm glad you have one issue fixed! As for your current issue...if you are trying to plot 2 separate lines in the same plot it likely will return an error. If you want, copy and paste your script that you are using and I may be able to better answer your problem!

@@OMAR-ce1bj I just created a new video that I think will provide the answer you are looking for! ruclips.net/video/IKoh4oiKtsg/видео.html

Thank you for the feedback! The short answer to your question is yes! It would take some adjustments in the file that you input, or the use of dplyr to arrange your data correctly and then facet_wrap within ggplot2. I will try to make a video soon to address this :)

​@@ecomattp Thank you so much for the reply! Looking forward to all your upcoming videos :D

Thanks for the feedback! I will see about making that video soon :)