Alternatives to ordination in R: Visualizing community change relative to a specific point (CC207)

[@thaisamoro]

Great video! Could you please clarify why filter (samples < names) is needed?

[@Riffomonas]

Thanks for watching! That gets you the lower triangle of the distance matrix so we don’t double count the same distance

[@wbsimey]

Thank you! Incredibly helpful. Have you made any vids on functional analyses? I have 16S PE data on twins and BugBase is not working for me.

[@Riffomonas]

Thanks for watching Brian! Unfortunately I haven’t done any videos with bugbase or metagenomics

[@scenarios911]

Hello sir, it's nice to see new graphs, but how is it better than ordination like in PCoA we can get same clusters graph even with more information like with each axes we can get percentage as well.

[@Riffomonas]

It's vastly superior to a single ordination because you aren't forcing your audience to mentally make this visual. Furthermore, I plotted the actual distances, not the distorted relationships between points that you will get with ordination.

[@yaqinguo8971]

Hi Pat, first thank you so much for your dedicated efforts. If I think twice, this jitter point plot definitely showed individual data points (which actually ordination plot also shows individual points) or variability, but this is not so obvious to understand the results. Probably I don't know how to interpret the results. If yes, please help me out (thanks in advance😊). But for example, if early group is significant different from late group, I believe this tells us late group is more cluster together than the early group. Or this tells us early group are more heterogeneous compared to the late group. But whether community composition is (significant) different or not between early and late group, I cannot tell. I suppose same (in your figure) could indicate this? But how large distance can we say the two community are really different from each other, there is no answer. or what we can know from the same data point? At least, with ordination plot, I can see how those two groups are separated and combined with statistical test (Adonis), I can say these two groups are different or not. But I think Ordination plot is only suitable for one factor, if you have more than one factor, each factor has more than two levels, it is disaster (they cannot separate well in the ordination plot). Above all are my thoughts. I look forward to your feedback. Because I want to use jitter plot, but I don't really know how this can tell me whether beta-diversity is different or not? Thanks so much!

[@zzayuzawa9289]

I love your videos. My question is how do I catch up with your videos? where should I start as a beginner?Thanks for the awesome work you do

[@Riffomonas]

Jump in anywhere! Watch enough and I'll start repeating myself at some point 🤓