A question is raised in my mind, then I need to take advantage of your idea and experience.
To draw an acceptable and well-analyzed taxa bar plot in terms of science in Qiime2 or any software package, should I essentially and scientifically pass the alpha-, beta diversity and rarefaction steps before? Can I publish unrefined reads/taxa bar-plot without scientific error and any risk?
One more question crossed my mind before creating the topic. Should I use unverified reads for alpha and beta diversity? What if I use ratified reads for them?
It depends on what you want to do. Taxa plots are pretty descriptive and give you a sense of whether or not things look right or wrong at a first glance (although it doesn't guarentee things are right). Alpha and beta diversity are community level descriptions.
Its rare that I believe papers that show a barchart and don't show diversity metrics. I have trouble believing life happens in stuff present at > 1% relative abundance and for me, interesting tends to be lower level. You might be working in a less complex system, and so it might be true there that interesting stuff happens at a higher level. If I were reviewing the paper, I would project reject or send back for major revision because they don't have enough evidence.
I'm going to recommend that you do a quick forum search on this one, and, if you can't find your answer, post it to a new topic.