Hi @ninaxhua,
You should for sure run DADA2 separately on each run as you suggested but I wouldn’t merge the tables just yet. Using 2 different regions will give you different sequences for the same potential taxa simply because you are looking at a different region. This as you can imagine can create false separation on a PCoA plot and can have confounding effects on further analysis. The 2 options that come to mind are
a) Treat the 2 tables separately all the way until you assign taxonomy to them, then merge the tables.
b) A better and more powerful approach would be to use the fragment-insertion plug-in described here which is specifically designed to deal with combining data with different regions.