Good afternoon!
I realize there is a post similar to this Beta-rarefaction diversity matrix export or mean distance?, but was hoping bringing it up again to see if anything has been changed since or could be produced. I have been using the script below:
qiime diversity beta-rarefaction
--i-table feat-table.qza
--i-phylogeny 2022.10.phylogeny.asv.nwk.qza
--p-metric weighted_normalized_unifrac
--p-clustering-method upgma
--m-metadata-file metadata_mouse.txt
--p-sampling-depth 100
--p-iterations 100
--p-correlation-method spearman
--p-color-scheme RdYlBu_r
--o-visualization Visualization/JK_w_norm_unif.qzv
My samples are so low biomass and are fairly unevenly sequenced, so I don't like to rarefy and loose samples or information. The plug-in above has been wonderful in visualizing the variability that can come from my samples. I appreciate the jackknifing technique since, as stated in multiple papers, rarefying once can lead to incorrect interpretation. I was wondering, though, if there was a way to use a "mean distance matrix" to perform the rest of my diversity analyses with. Essentially, I would like a matrix constituting of the means of each sample centroid displayed in the beta-rarefaction output PCoA so that I could do statistical analysis. Theoretically, the feature table has already been rarefied 100 times and distance matrices have been made for each iteration. I get that the code outputs temp files since it would be too much to store, but could it not output one final matrix that is a mean of all 100 matrices, for example?
As stated in the topic linked above, I had also thought of manually creating 100 rarefied tables and then making distance matrices for each of them. That would essentially mean I would have to put my computer through running the exact same process all over again (and my poor computer has already been a trooper, I don't want to strain it with more hours of work) just to get to the data one step before the PCoA output. I wanted to know if there was now a qiime2 plug-in that would do that for me or if beta-rarefaction could add that as an output.
Since I have already made the plots, is there any way to extract the distance matrix that made the PCoA plot? I would like to be able to make heat-maps and run stats on them.
Thank you so much for your time! I really appreciate the work that's been done to make these plug-ins in the first place and I'm excited to implement them in my research.
Marlyd