How to choose wavelengths for MAD
Phasing by SAD requires density modification to break ambiguity, therefore it works better with high solvent content crystals. Also, data to 2.5Å or better are required and high resolution native data for phase improvement and extension by density modification are a great advantage.
MAD instead: usually wavelengths are precisely selected, and data are redundant and precise, therefore data are stronger than in SAD. This means that MAD will work also for low-resolution phasing, and especially it does not depend on density modification to break ambiguity, so it can work also with low solvent content.
Anomalous merging R (Ranom)
A ratio Ranom / Rpim > 1.5 is a good indication of the likelihood that substructure solution based on anomalous difference will succeed . For Rpim see here, while Ranom is defined as:
RedundancyAnomalous scattering violates Friedel's law: reflections (h,k,l) and (-h,-k,-l) have different intensities, and this difference (that gives the name anomalous difference) translates in a low redundancy (the ratio between the number of observations and the number of symmetry-merged reflections) because of a loss of symmetry in the diffraction pattern. Therefore when collecting MAD data we need at least twice as much data per wavelength than for a native dataset.
Because of the need to collect more data, radiation damage becomes a serious issue in MAD data collections. Therefore, we need to expose less than for a native dataset for the same crystal, cutting exposure times by at least a factor 6.