I came out the cave with no programming background and understood this. On the plus side, you also find your way through the road, which is how I do pretty much everything. This is how everything should be taught. Liked and subscribed! By the end, I'm going to be able to write and read one more language: R. Thanks for putting this together and teaching one BioChem engineer from the third world statistical analysis using R. Cheers!
Recently found out about your channel and your tutorials are amazing. Short, easy to follow, practical and fun. Clearly a lot of work has been put into those videos. A huge thank you
Did we ever discuss the pros and cons of PCoA (MDS) and nMDS? Would love to see input from you on the topic as well as how generate PCoA using a given data.
I feel like I may have made an oblique reference at some point. The files used in these videos were generated using mothur. I prefer nmds to pcoa because the nmds visuals represent more of the variation in the data than pcoa will for the same data. The downside of nmds is that the axes don't really represent anything whereas the pcoa axes are interpretable. I see ordination as a tool for visualizing data rather than as an analysis on its own. And... I think ordination is generally very over rated as a visualization approach :)
Actually, i am beginner of microbiome analysis this lecture is so useful ! If possible, could you show how to generate raw data of nmds in this lecture? Thank you!
Hi there! You can see how to generate the axes files by following the mothur MiSeq SOP. The command used to run nmds is nmds. mothur.org/wiki/miseq_sop/
Thanks for the videos, have you used phyloseq object somewhere in the analysis? It is quite useful. In your GitHub: raw_data, there are the necessary requirements to do so.
Hi Wisam - thanks for watching! Nope, I've never used phyloseq. I know people like it, but I'd prefer to teach from fundamentals like ggplot2, dplyr, and base R so that people can take their skills to do anything.
@@Riffomonas thanks for your reply. I think I will make it via your data and refer to you of course. Have you removed the unclassified taxa from the original dextar taxanomy table?
Pat, a simple question. How do you manage to see the figure output instantaneously. Is there a shortcut key in RSTUDIO or windows 10 that we can use to generate the pop-up figures instatneoulsy as you show in your videos. In normal cases we have to go to the folder the figure is saved and open it from there. Thanks for the help.
I use ggsave to save the figure to a file (see the last line of my scripts). When the figure gets saved the figure updates. I also do a little editing to make it seem a second or two faster than it really is
I feel as though I missed a video - do they go in order from CC077 to CC078? where is the code coming from that you reference here? I can see the blog post and what to paste but am coming up with several issues. Can you direct me to the video on where to find step by step to get to this point?
Thanks for watching - sorry for the troubles. Your error is what we typically see when an object hasn't been created yet. I wonder if you might have missed running metadata
Please, i have been trying to run your data, but i keep up getting the error below, how can i rectify it Error in `inner_join()`: ! Can't join on `x$sample_id` x `y$sample_id` because of incompatible types. ℹ `x$sample_id` is of type >. ℹ `y$sample_id` is of type >.
Thanks for watching! You'll need to either read the sample_id column in as a character or use mutate(sample_id = as.character(sample_id)) on that data frame
Thanks for watching - check out the starting code which is available at riffomonas.org/code_club/2021-03-08-scatterplot. That has the code to generate metadata_nmds
You mentioned the limitation of making the species name italic. I find the ggtext package ( github.com/wilkelab/ggtext ) very useful when I find myself in the same spot as you here, and wonder if this could have helped in your legend formatting since you didn't really seem happy with the labelling of your levels. Anyways, following CodeClub for a while and still learning new things every episode, keep up the excellent work, really enjoying it.
What do you think of my rubric? Anything you would add? Any thoughts on other ways we could improve this figure?
very helpful!
I came out the cave with no programming background and understood this. On the plus side, you also find your way through the road, which is how I do pretty much everything. This is how everything should be taught. Liked and subscribed! By the end, I'm going to be able to write and read one more language: R.
Thanks for putting this together and teaching one BioChem engineer from the third world statistical analysis using R. Cheers!
Recently found out about your channel and your tutorials are amazing. Short, easy to follow, practical and fun. Clearly a lot of work has been put into those videos. A huge thank you
Hey Cyril- thanks so much! I really enjoy making the videos and it’s always great to hear people are getting so much out of the videos.
Thank you so much. The nitty gritty of the labels and legend position were helpful
This was superbly done and very instructive!! One of your best ever sessions! Many thanks 🙏
Thanks! Keep watching 😂
Just starting after meeting your student on the flight to Microbe 2022... Big thanks!
Fantastic! Glad to have you as a viewer
Thanks for this content! Just what I needed to get my publication done! Again thank you SOOO MUCH!!!
Hey Alejandro - wonderful to hear! Good going :)
Thanks for tutorial!
My pleasure! Holler if there's any topics you'd like to see me cover in the future
I'd go through anything you put out. This video was quite basic for me, but I know that I would learn something. I was right.
That's very helpful! Thank you!!
Did we ever discuss the pros and cons of PCoA (MDS) and nMDS? Would love to see input from you on the topic as well as how generate PCoA using a given data.
I feel like I may have made an oblique reference at some point. The files used in these videos were generated using mothur. I prefer nmds to pcoa because the nmds visuals represent more of the variation in the data than pcoa will for the same data. The downside of nmds is that the axes don't really represent anything whereas the pcoa axes are interpretable. I see ordination as a tool for visualizing data rather than as an analysis on its own. And... I think ordination is generally very over rated as a visualization approach :)
Actually, i am beginner of microbiome analysis this lecture is so useful ! If possible, could you show how to generate raw data of nmds in this lecture? Thank you!
Hi there! You can see how to generate the axes files by following the mothur MiSeq SOP. The command used to run nmds is nmds. mothur.org/wiki/miseq_sop/
Thanks for the videos, have you used phyloseq object somewhere in the analysis? It is quite useful.
In your GitHub: raw_data, there are the necessary requirements to do so.
Hi Wisam - thanks for watching! Nope, I've never used phyloseq. I know people like it, but I'd prefer to teach from fundamentals like ggplot2, dplyr, and base R so that people can take their skills to do anything.
@@Riffomonas thanks for your reply. I think I will make it via your data and refer to you of course. Have you removed the unclassified taxa from the original dextar taxanomy table?
@@Wisam_Saleem Nope, all the original taxa are there
Pat, a simple question. How do you manage to see the figure output instantaneously. Is there a shortcut key in RSTUDIO or windows 10 that we can use to generate the pop-up figures instatneoulsy as you show in your videos. In normal cases we have to go to the folder the figure is saved and open it from there.
Thanks for the help.
I use ggsave to save the figure to a file (see the last line of my scripts). When the figure gets saved the figure updates. I also do a little editing to make it seem a second or two faster than it really is
I feel as though I missed a video - do they go in order from CC077 to CC078? where is the code coming from that you reference here?
I can see the blog post and what to paste but am coming up with several issues. Can you direct me to the video on where to find step by step to get to this point?
this is the error code I keep getting:
Error in inner_join(metadata, nmds, by = c(sample_id = "group")) :
object 'metadata' not found
Thanks for watching - sorry for the troubles. Your error is what we typically see when an object hasn't been created yet. I wonder if you might have missed running metadata
Hi Thanks you so much Please could you explain about how pick OTU from NGS data
Thanks is for watching! I’d encourage you to check out the MiSeq SOP at mothur.org to see how we do it
Please, i have been trying to run your data, but i keep up getting the error below, how can i rectify it
Error in `inner_join()`:
! Can't join on `x$sample_id` x `y$sample_id` because of incompatible
types.
ℹ `x$sample_id` is of type >.
ℹ `y$sample_id` is of type >.
Thanks for watching! You'll need to either read the sample_id column in as a character or use mutate(sample_id = as.character(sample_id)) on that data frame
Hi, very good video. I am wondering how to shift figure 90degree?
You could flip the x and y aesthetic values or you could use coord_flip
Hi there, it is really amazing video. However, I keep getting the error "metadata_nmds' not found"
Thanks for watching - check out the starting code which is available at riffomonas.org/code_club/2021-03-08-scatterplot. That has the code to generate metadata_nmds
You mentioned the limitation of making the species name italic. I find the ggtext package ( github.com/wilkelab/ggtext ) very useful when I find myself in the same spot as you here, and wonder if this could have helped in your legend formatting since you didn't really seem happy with the labelling of your levels.
Anyways, following CodeClub for a while and still learning new things every episode, keep up the excellent work, really enjoying it.
Agreed! I did use ggtext in a previous episode and will use it again in a future episode
Hi I would like to get your opinion on my project and graphs. please let me know how to reach out.