I created a sequence manifest and am demultiplexing. There are two batches of sequences (batch A and B). I'm doing quality control on them separately and then merging them later. However, after demultiplexing each batch, the histograms in the qzv file looks different despite using the same code for each (see code below). Batch A has a single histogram but Batch B has two histograms describing forward and reverse reads with the same exact numbers. Why does this occur? paired-end-demux_A.qzv (294.1 KB) paired-end-demux_B.qzv (315.6 KB)
Judging from the provenance in your visualizations, you used two different versions of QIIME 2 at two different times — one visualization was generated with 2020.2, and the other with 2020.8.
It looks like several adjustments were made to the visualizer in that time to add support for reverse reads etc, so the difference in appearance is just related to these changes. These changes are mentioned in the release notes, though they are not too specific, but this is a good way to generally see what changed, e.g.:
That's the distribution of per-sample read counts for each direction. The fact that they match is a good thing! It's exactly what you want to see — paired-end reads should match in their paired samples.