Want to use Vegan::capscale to calculate and plot distance based- redundancy analysis (db-RDA) and add correlation cut-offs in the results

[Tahsin-Khan]

I am quite new in using R.

I have a dataset with relative abundances of Genomes against their COG-ids. It is a large data frame with 4214 rows (COG-ids) and 1313 columns (Genomes).

https://drive.google.com/file/d/1byXVMMWYcCgVwqRRLTZV6UJf7ejCBucE/view?usp=sharing

I would like to

make a Bray-Curtis distance-based PCoA plot with my data which will show clusters of similar COG-ids based on their relative abundances from the data frame.

do a distance-based redundancy analysis (db-RDA) to find out which COG-ids are driving the main clusters to set apart from each other.

As there are 4213 COG-ids, I want to use some kind of correlation cut-off in my analysis, so that I get top 20% COG-ids responsible for driving the functions rather than all 4213 in the RDA plot, which will make it look very clumsy.

So far, I have tried to use "capscale" function in Vegan using distance bray to do this analysis but have been getting the error "Error in cbind(x$CCA$v, x$CA$v) : number of rows of matrices must match (see arg 2)" when I am trying to plot the RDA.

With my very small knowledge, I cannot comprehend this problem. Would be very grateful with some help on this problem.

Thanks in advance.