Performing principal coordinate analysis (PCoA) in R and visualizing with ggplot2 (CC186)

[@shillawolf4686]

I am working on a thesis project, and i had to run pcoa. But i had no clue as to how and your video really helped me to understand what i should do and why these steps are taken. Know that your videos are helping a lot of students out there to better understand how to perform statistical analysis and we are very gratefull!

[@andreamaritin2856]

Thank you very much! I was strugling with r projects in my master's degree and you just saved my grades.

[@audreythomas4324]

I wish I had found this before 2am 😅this explains so much better than anything else I've seen

[@zmac18]

Just in my time of need! Thank you!

[@Riffomonas]

Wonderful - I’m glad you found it helpful!

[@cristianjaviermena6955]

awsome doc! thanks a lot. from argentina watching! and learning

[@absartalat]

Thank you for this. Looking forward to NMDS lecture.

[@Riffomonas]

Thanks for watching Absar!

[@russtin1]

My heart sank when you hard coded those numbers. I should have had more faith in you.

[@Riffomonas]

Ha!

[@xiaojunslefteyebrow]

8:56 how can I do this step to get positive eig values in Python? I did the rest of the project I’m working on in Python

[@xiaojunslefteyebrow]

please?

[@Riffomonas]

I don't know python

[@kattiacristina]

Great!! Thanks for this video

[@Riffomonas]

My pleasure! 🤓

[@haraldurkarlsson1147]

I agree with base R plots vs ggplot2 but still it is good to know. Many of the packages out there still use base R - especially the older ones. :(

[@djangoworldwide7925]

OMG i got sick when you hard coded the axes :OO thank god you fixed it :P

[@jiayunzou9867]

Thanks for this wonderful video!! If we need to extract the first two axes for analysis of microbial community composition. I was wondering which one is better: PCoA or NMDS? Really looking forward to your reply, thanks a lot!

[@Riffomonas]

Probably PCoA

[@jiayunzou9867]

@@Riffomonas Thank you for the timely reply, it helps!

[@rojalinbiswal6723]

when I am trying to plot getting error so how can I find your data ie., sample data to run the same code ? please answer .

[@Riffomonas]

You can find the link to GitHub at www.riffomonas.org/code_club/2022-02-10-pcoa

[@alessandrorosati969]

How is it possible to generate outliers uniformly in the p-parallelotope defined by the coordinate-wise maxima and minima of the ‘regular’ observations in R?

[@Riffomonas]

Sorry but I'm not sure I understand what you're asking. It's likely beyond what I was trying to teach in this episode

[@valentinberrios5927]

Hey thanks so much!, It captures my attention the fact that there is lower explained variation by the two first axis in comparison with PCA analysis. For my understanding , it is more acceptable when the sum of the explained variation by the two first axis is greater than 50% in PCA. It seems like it is a tendency to get lower explained variance in PCoA isn't it ?

[@sven9r]

lovely! How do you feel about PCoA vs NMDS in metabarcoding studies? I think the question of which ordination is appropriate for the given data, is one of the most common source for mistakes in modern ecology.

[@Riffomonas]

I prefer nmds but I’m only using them to make pretty pictures

[@sven9r]

@@Riffomonas ty looking forward to the next episode :) have a nice weekend

[@manueltiburtini6528]

Don't you need to perform Bartlett sphericity test or KMO test before? What's about triangle inequality and Calliez correction for negative eigenvalues?

[@Riffomonas]

When done with a distance like Bray-Curtis those are not performed

[@priyankamishra3632]

how to plot a pcoa for pathways instead of showing the samples?

[@Riffomonas]

Thanks for watching! You’d have to have a similar matrix but where the pathways are on the rows in place of the samples or in the columns in place of the taxa. What you do depends on what you want each point to represent

[@priyankamishra3632]

@@Riffomonas Thank you for the clarification. I'll try in this way.

[@giovanniwidmer29]

Pat, at the risk of sounding professorial, it's principal, not principle. Thanks for the excellent education material!

[@Riffomonas]

Gee whiz. Thanks 🥴

[@rajaldebnath6007]

Rival software package: QIIME. 😂

[@Riffomonas]

😉