I'm not familiar with Drosophila cells, but we use both MACS and FACS for mouse and human hematopoietic cell isolations. I routinely use positive selection, passing the cells over two columns. My initial population is 1%, after the first column 10-30% and after the second column >95%.
FACS in my hands is usually more pure (easily >99%). Since a more pure population is preferred for RNASeq (so that impurities don't influence your data) I would recommend using FACS. A reason for using MACS would be that it is more gentle on the cells, so if your cells don't survive FACS sorting, MACS might be an alternative, but always check the purity of the selection afterwards.
Thank you Martijn Brugman. As isolation using FACS would be more harmful to the cells, do you think it induces a stress response that can change the Regarding timing, which method is faster?
Time wise, it depends on how many cells you have and what starting percentage you need to start from. FACS takes me about an hour for staining and washing, and then 1-2 hours for sorting.
MACS takes about an hour for staining and each column also 30 minutes, for a total of 2 hours.
Even with the possibility of harming your cells a little I would still prefer FACS because of the superior purity. We have had good results for RNASeq but also hybridization arrays for peripheral blood cells, bone marrow cells, MSC, so FACS is applicable for many cell sources. If you have very large cells, such as MSC, you might want to use a slower sort speed and a bigger noozle (e.g. 100 µm instead of the 70 µm standard)