I ran B-diversity analysis on my data. And upon looking at the unweighted unifrac plot, I noticed that the data is segregated in 2 groups as seen the image below, but this grouping is not explained my any of the component of my metadata. I wanted to know if there is a way to find what factors are responsible for this grouping?
I performed a B-diversity analysis on my data. From the unweighted UniFrac plot plot,, as shown in the image below, it is evident that the data is segregated into two distinct groups. However, none of the components in my metadata can explain this observed grouping. I am interested in exploring if there is a method to identify the factors responsible for this grouping.
1 Like
Hello!
In that case I would go back to the metadata file and try to add as much data as I still can collect for the dataset.
It can be library preparation, extraction, sequencing run, experimental run and other factors, then recreate PCoA to see if some columns explained such separation.
Best,
2 Likes
Hi Timur,
Thanks for your suggestion. In this case, all of the samples were processed together, from extraction to sequencing. I have carefully examined the metadata, including parameters such as sample collection dates, geographical locations (zip code, state, region), and more. However, I couldn't identify any other factors that could explain the observed separation in the plot.
My specific query is regarding the separation along Axis 1 in the plot above. I'm interested in finding out which factors may be contributing to this separation. Is there a method or approach that can help me determine the factors responsible for the variation along Axis 1?
Please let me know if there are any additional analyses or techniques that you would recommend to investigate this further.
Thank you for your guidance and assistance.
Thanks,
Gaurav
You should be able to export two lists of samples for the two groups you see on Axis1. Finding out what 'pushed' samples into these two groups is up to you. 
An anecdote from my work: we were doing a diet swap study in mice. 
We thought our important categories were diet and genotype, but the two buildings storing the mice was the largest effect.
We published it anyway and it's been cited 90+ times: Influence of early life exposure, host genetics and diet on the mouse gut microbiome and metabolome | Nature Microbiology
If you knew what you were going to find, it wouldn't be called research 
1 Like