Possibly similar but you can still take a look at this thread here which should be ideally similar to what you are asking. In any case you can ise GOrilla for GO enrichment when you can find enrichment and also number of gene counts that are enriched for each GO terms , gives you a number for farther filtration. Or you can always select the GO terms based on FDR threshold and then try to plot GO term over-representation using ReViGO for visualisation. [revigo is out of context but still if GO over-representation might be your downstream need]
Above way are not based strictly in terms of R but if that is what you want then GOstats or TopGO or GO.db can be something you can take a look.