qiime diversity core-metrics-phylogenetic

[Angeliki (Angeliki t)]

I would like to ask about the resuts of qiime diversity core-metrics-phylogenetic. The resulting metrics of diversity a or beta are calculated based on which taxonomy level? Can we change the level of taxonomy used?

Thank you very much!

[timanix (Timur Yergaliyev)]

Hello!

Phylogenetic metrics are calculated based on ASVs ' phylogenetic distances based on the tree one should construct before running core-metrics.

Best,

[Angeliki (Angeliki t)]

I would like to use ALDEx3 for a statistical analysis of the groups produced in the pcoa resulting from the bray curtis in core metrics. Which table should i use? Usually, I use the counts of reads assigned in each taxonomical level ex:genus, by the classifier for statistical analysis, but in this case they were not used to produce the distance matrix because it was produced based on the tree. Now which table can i use?

Thank you very much

[timanix (Timur Yergaliyev)]

Hi @q2-mods
I think I need help with this one. I am not familiar with applying ALDEx3 on the PCoA output. Any tips here?

PS: How can one whisper now? Can't find it anymore.

[Angeliki (Angeliki t)]

not in the pcoa output, using as groups the groups from a clustering in the pcoa output. I am saying that i need to use the same data from which the pcoa was calculated, meaning the distance matrix of beta diversity that was calculted on ASVs and not on the taxonomic level ex: genus. The distance matrix is different when calculated for Family, Class, or Genus and should be also different when calculated fro ASVs that have not yet been put into taxonomy. So i need that matrix that the core metrics plugin uses. Is it the matrix that i can export as TSV from the biom table of feature table? Is this enough? Because it also uses the tree, how can give ALDE the right matrix of read counts from which i decided how many groups and of which composistion i give to it to use for the comparison it calculates?

[timanix (Timur Yergaliyev)]

Hi @Angeliki
Thank you for clarifications!
I think that there are a lot of things that get mixed up.

@Angeliki

not in the pcoa output, using as groups the groups from a clustering in the pcoa output.

Do you want to compare different groups present in your metadata file (option 1), or did you notice distinct clusters in your PCoA output (not in the metadata) and want to compare them (option 2)? If option 1 is correct, you can just use your metadata file. If option 2 is correct, you can add this information (e.g., Cluster1, Cluster2,...) to the metadata file and use it for the test.

@Angeliki

I am saying that i need to use the same data from which the pcoa was calculated, meaning the distance matrix of beta diversity that was calculted on ASVs and not on the taxonomic level ex: genus. The distance matrix is different when calculated for Family, Class, or Genus and should be also different when calculated fro ASVs that have not yet been put into taxonomy. So i need that matrix that the core metrics plugin uses. Is it the matrix that i can export as TSV from the biom table of feature table? Is this enough? Because it also uses the tree, how can give ALDE the right matrix of read counts from which i decided how many groups and of which composistion i give to it to use for the comparison it calculates?

Aldex3 will not accept the distance matrix as an input - you need feature table (feature counts) and metadata file. You should have a feature table at ASV or any taxonomy level if you calculated diversity metrics - it is the same table as used for the input to it. Matrix is calculated by the core-metrics plugin, and it is not the biome file. The biom file is actually a feature table with ASV counts, while the matrix is a tsv file that contains sample-to-sample distances.

In short - you can use the same table for both core-metrics and Aldex3 test. Aldex3 can't accept the distance matrix. If you need ASV counts, you can export feature table as biom file and then convert it to the tsv. You can find example commands on this forum by using "search" field.

Best,