Dear John, it is amazing to note how frequently you force the readers to think about something out-of-the-box. Thank you.

Now, why do we need crystallography? If the answer is "to know the conformation-function of proteins", then ideally we should be conducting crystallography for all existing proteins, which is not possible. Besides, I have developed a feeling that crystallography is half-science and half-luck; you would never be sure if a certain protein that you are trying would be possible to make in a bacterial strain; if it does, whether it could be purified satisfactorily; if it is, whether it could assume a stable conformation; if it does, whether it can be crystallized or not. With so many ifs and buts, I would opt for other methods for structure analyses. Besides, molecular, structural and informatics biologists must now come together to develop "reliable" sequence-based conformation modelling.

But, clearly, we do not have adequate number of sequence-structure combinations elucidated thus far that could pave the way for informatics-based modeling. Definitely we need sustenance of the crystallography research for a few decades more. But, the only thing that we can say is: it is an unfortunate situation. I guess several fields of science--which helped humanity develop immensely--will become obsolete pretty soon.

Decreasing funding for macromolecular crystallography is a major problem, I agree. Crystallography has been a critical tool for understanding molecular function and is likely to remain so for decades, despite dramatic improvements in other methods such as cryo-EM. The end of the Protein Structure Initiative in the US will put a large group of crystallographers back into the general competition for individual-investigator research grants and make it considerably more difficult for US investigators to obtain funding. In my view structural biology is so important, and as John pointed out, the results so enduring, that it should be strongly supported for a very long time.

What is the likely outcome of the current trend? Though structural biology may have a special importance, its situation is not very different than other biomedical fields. The size of the research community is larger than the amount of funding can adequately support. In the absence of additional funding, the number of investigators as well as the size of grants and the number of students, postdocs, and technical staff will decrease. I hope that this decrease will be orderly and that it will not come at the expense of the newest entrants into the field. If we are lucky, then a new stable but smaller community may emerge with adequate funding for those who are members.

I understand the need to fund all areas of the research community, but think the need for structures in the face of ever faster sequence data production is growing faster than the number of crystallographers we have currently.  So this is at odds with cutting back the number of people able to efficiently solve structures. 

In a stressful time for all of NIH funded science, let me broaden the question to all of biology in general.  Every field in biology wants to generate enormous data, be it atomic structures; proteomics of cells, organs and whole animals; knockout mice-knock-in mice in the thousands and in various genetic backgrounds as well; genomes of hundreds of organisms and humans; document every post translational modification; this list just goes on. Every grant has justifications too; basic knowledge, cure cancer, makes a vaccine, etc. So how does a NIH panel decide what to fund and which areas to prioritize? (How economics shapes science by Paula Stefan, Georgia State University, is a good read)

Historically, biological crystallography brought mathematical rigor and chemical precision to biology on a large scale (Apologies to Ronald Fisher and population genetics). But the rapid evolution of precision genetic techniques is making crystallography obsolete in the academic sphere, but it will survive and thrive at drug makers. Take for example gene knockout techniques, earlier you needed some human disease or condition to identify the gene of interest. It was tedious to sort out the mechanistic details of disease development. Now, mice are the new E.coli, one can simply order knockout mice and study everything associated with the condition. The information content in terms of physiology, generated by mice studies is immense. You can have site directed mutants in mice too. Mouse models of disease are immensely useful for physicians and scientists (they have their limitations too). Try convincing a doctor that a crystal structure is better than an animal model to understand a disease process; the animal wins by a walkover.  So the question is why do we need protein structures on a large scale? (we are talking in the thousands). The biggest life savers of the last century, antibiotics and vaccines did not need crystallography!

Chemical understanding of biology needs structures. But Biology has to be understood from multiple angles and not just the reductionist manner. My point is that a host of other fields are also equally important and need funding. Let’s put ourselves in the role of policy makers and the problems look very different (which area will give the biggest bang for the buck). Historically, good economic conditions have lead to investment in the sciences not the other way around. The steam engine did not need thermodynamics either. That’s my thoughts.