RNA-seq: thresholds for quality control & batch correction

Question

[GenomicsNewbie]

After mapping the reads from TopHat, I see that I should check various quality control metrics. Most widely used quality control metrics include sequence quality, sequencing depth, reads duplication rates (clonal reads), alignment quality, etc. Can you teach me what these metrics tell you about? Also, what is a good threshold to use to say "good quality"? Following up on this, how do you deal with those "bad quality" reads?

Another question is, how do you identify batch effects from your mapped reads? I assume that this is a separate step from QC?

[swbarnes2]

https://twitter.com/lpachter/status/937055346987712512?lang=en

"Please stop using TopHat"