MiSeq output quality distribution.


Even though these graphs are produced by qiime2 (2020.8), the question is about the sequences themselves or the sequencing process, not the plugin output.
When performing 16S V3-V4 analysis on two different sample cohorts(different groups of humans) in which the metagenome samples are collected from different environments (Cohort 1: Environment A, Cohort B: Environment B, unrelated), i noticed kind of a trend after performing adapter trimming and the following demux summarize summary. I am attaching the pictures below. All the samples from both cohorts were sequenced on the same Illumina MiSeq platform with 2x250 PE reads and 341F/805R primers.

Regardless of the overall quality, which obviously is different between the cohorts, i have highlighted (in paint) the overall quality drop at around 80bp for forward reads in both cohorts and also the drop at 80 and slight rise at 100 for reverse reads in both cohorts.

My questions are:

  1. Has anyone else, who has used this sequencing platform before also seen the same trends?
  2. Could this be just a product of the nature of illumina sequencing process(e.g. new reagents flooding in at 80 cycle or something to that matter)?
  3. Could this be a PCR related issue?
  4. Maybe a bad batch of sequencing chips?
  5. Maybe it is related to V3-V4 regions?

Any opinions or suggestions are welcome. Asking out of pure curiosity.