I came across a few posts in the forum about how to filter-out chloroplast and mitochondria from a given feature-table. That is, by setting
--p-exclude mitochondria,chloroplast from
qiime taxa filter-table. However, I’ve got these samples and they happened to be over 90% represented by chloroplast and mitochondria (no pre-treatment against those RNAs). I am now intrigued if that is how it often happens, and I’ve found some literature pointing to very high chloroplast sequences in 16S samples, but I would like to know if there is any way of checking if any of these enormous ammount of sequences are not chloro/mito in a sense that I would reclassify them.
Any help on that?
Yikes! That is very high indeed. The presence of mitochondria and chlorplast can be inflated depending on your sample source , DNA extraction methods, and contamination. Though, 90% is certainly much much higher than I’ve ever seen, even though I’ve generally dealt with gut/fecal samples I can’t recall any other sample sources with such high abundance of those.
As for misclassification, it is I guess in theory possible. You could try blasting a few of those features that are being identified as chloroplast or mitochondria and see if there are other taxa similar. How long are the reads you are working with and what is the target region?
Thanks for the suggestion. I am blasting those for a deeper look.
Median size is 277 nts, single-end targeting V3 (341F).
Yes, after all checkings it does’n seem to be something other than chloroplast. Will filter it out and follow with the filtered samples.