Managing resources is important to your success in obtaining useful results. With too much effort, you may wind up with redundant results. With too little, you may get nothing or poor results which do not optimize after weeks or months. It is important to recognize some truths, which are discussed below.
You should be primarily concerned with the quality of sample going into crystallization plates. If you are able to spend a little extra time or resources, you can avoid weeks of evaluating failed experiments and repeating previous attempts. Get your sample to be stable and pure before you set up trays. In the long run being meticulous will accelerate your pace.
Pretty crystals are not the same thing as useful crystals. It is a common occurrence for a sample to crystallize in multiple conditions, some of which diffract better than others. Diffraction quality cannot be determined by looking at sharp facets or anticipating a certain morphology. Remember that you are screening with amounts of protein and crystallant that are fairly random. These will not reach their ideal size and shape immediately. Even with optimization, certain conditions will diffract far better than others. You need to get feedback from an in-house source to decide where to put your efforts. Without this you will be guessing. By investing a small amount of time to evaluate crystal quality, you can get feedback between synchrotron trips and make informed choices.
SETTING UP ENOUGH EXPERIMENTS
Your sample will not behave like “similar samples”. A single residue mutation can have a dramatic effect on what type of crystallant is needed to obtain crystals. Unless you are screening a sample that is known to crystallize in most conditions, you need a broad approach. A reasonable approach is to set up from 3-5 broad trays for a new sample. You also will have more success by attempting multiple constructs at the same time. A common pitfall is the trial and error approach, in which attempts are made with single constructs, fail over the course of weeks/months, and are abandoned for another construct. This approach leads to lengthy projects and low yield. By setting up a few trays for several variations of a sample (mutation, truncation, ligand present), you will save a great deal of time and have a much better chance at getting a result.