For comparison/validation, I suggest a stacked barchat of most abundant (read count) taxa across all samples.
x axis = sample
y axis = read count OR otu count, coloured/grouped by taxon, pooling less abundant taxa as "other".
x-axis may reasonably hold up to ~100 samples in one screen(?) - BUT: sample id (labe on X axis should be viewable, to be able to ID outliers and refine/correct dataset based on info from the chart). Paging will be needed.
For comparison/validation, I suggest a stacked barchat of most abundant (read count) taxa across all samples. x axis = sample y axis = read count OR otu count, coloured/grouped by taxon, pooling less abundant taxa as "other". x-axis may reasonably hold up to ~100 samples in one screen(?) - BUT: sample id (labe on X axis should be viewable, to be able to ID outliers and refine/correct dataset based on info from the chart). Paging will be needed.
For inspiration https://www.google.com/search?sca_esv=559959589&rlz=1C5GCEM_enDK1026DK1026&q=taxonomic+composition+barchart+metabarcoding&tbm=isch&source=lnms&sa=X&ved=2ahUKEwiYs-jZsPeAAxUWbPEDHeGxA0QQ0pQJegQICxAB&biw=1496&bih=848&dpr=2