This entire process is one of sampling. You initially take a "sample" from your experiment. A part of that sample is going to be converted into a sequenceable "library". A small fraction of that library is what you sequence to get your initial data (so there is a lot left over). If you think that data is not enough you can always go back to your library and sequence more. As you do this process you will find that while you may find some "new" data (miRNA in your case) much of it you would have already seen in round 1 of sequencing. This is a process of diminishing returns as you keep sequencing more. If the library stops yielding new sequences you may need to go back make another experimental sample and repeat the process. This is the reason > 3 biological replicates are standard recommendation.