| Newsletter No. 391
|| February 4, 2019
ACA 2019, FEBS 2019, Protein Society 2019
DISCUSSION - Table 1
A perspective article "Against Method: Table 1--Cui Bono?" was recently published in Structure by Bernhard Rupp, pointing out that "Information contained in 'Table 1' is insufficient to evaluate or repeat the experiment, is redundant with information extractable from deposited diffraction data, and includes data items whose meaning is under increased scrutiny in the crystallographic community," and suggesting that "Direct and consistent extraction and analysis of data quality metrics from preferably unmerged intensity data with graphical presentation of reciprocal space features, including impact on map and model features, should replace 'Table 1'."
Do you agree with Bernhard? Please share your opinion and ideas with fellow crystallographers.
Dr. Lothar Esser: Commentary on “Against Method: Table 1 – Cui Bono?” B. Rupp Structure 26 (2018) 919-923. In his perspective Hofkristallrat Rupp argues essentially that ‘Table 1’ should be ‘relegated to the depths of supplementary material where it will be […] rightfully forgotten.’ Table 1 in brief refers to a tabulated summary of direct crystallographic data like cell dimensions and symmetry as well as derived characterizations of the presented molecular structure like root mean square deviations of bond lengths and angles from standard values. It is suggested that this information is neither useful to reproduce the experiment (presumably diffraction experiment), nor does it help to evaluate the quality of the final structure model and is in fact redundant because all values can be retrieve from databases easily. Table 1 could however be maintained if outmoded and woefully inadequate statistical values be replaced by 21st century quality indicators like CC* and enhanced graphic depictions of reciprocal space. This kind of unmitigated assault on Table 1 is not quite new and is to be understood in the broader context of an ongoing process of improving quality control and giving editors and reviewers meaningful tools to correctly evaluate a given new structure in the absence of having all data available for full-fledged inspections. All this introspection is meant to benefit the structural biologist whose work critically depends on having exact knowledge of the structure down to side chain rotamers, ligand positions and hydration of catalytic or other sites of interest. Being in this field for a while, I have yet to meet one of those beneficiaries – someone who works with structural data but does not know how to evaluate its quality. In conclusion, Table 1 would benefit, as everything else from a critical review and update to 21st century standards. The reader can only draw one conclusion equally applicable to my comments and Prof. Rupp’s perspective in response to the self-imposed question ‘Cui bono’: No one.
Dr. Mariusz Jackolski: "Dr. Xinhua Ji has encouraged discussion on the somewhat controversial proposition of Dr. Bernhard Rupp to get rid of Table 1 from published crystallographic papers. As much as I sympathize with Bernhard's crusade to wed "post modern relativism" from exact science, I am not sure that his aim at Table 1 is well advised. I am afraid that his proposition could have the opposite effect, i.e., could accelerate the demise of something that is perhaps not perfect but certainly very useful; in the end Table 1 would be replaced by just nothing. Bernhard is giving strong arguments to the opponents of any kind of Table 1 information in experimental papers. If we want to maintain high standard in crystallographic literature, we should work out a concrete (improved) alternative to Table 1 and start using it in practice."
Dr. Peter D. Sun: “I do not agree with the author of the perspective article in Structure. I think one has to ask who reads Table 1 and how useful it is in a crystallographic structure publication. Without giving too much thought, I think Table 1 is very useful to most non-crystallographic trained but structurally tuned readers—which probably makes up the majority of the scientists who are interested in crystal structures. Majority of them do care about resolution and a few also care about R-factors. Table 1 serves primarily the purpose of announcing the birth of a “new” structure with certain attributes to allow general public to assess the quality and compare the current on with any other structures they know of. In the past, crystallographic community have relied on Table 1 to evaluate a structure for its correctness. It is true that the information in the table is insufficient to repeat the experiment and in many cases indifferent to minor errors in structures. However, in this day and age and with the exception of few, most of crystallographers read structural publications for new scientific insights and structural information, not with a purpose to validate and repeat the published structures. This is mostly due to the software development in our field.
Secondly, I would caution to use highly specialized statistics for data quality representation. Even though such representation, such as unmerged intensity data with graphical presentation of reciprocal space features, may indeed reflect more accurately the state of a refined model, they are less comprehensive to our audience, and thus less useful. In fact, it is not clear how many crystallographers understand the errors and their propagation in our models.
In anycase, my point is a structural publication is best regarded as an illustration to a new consumer product. It is for the interest of general audience. For a few specialized experts, then again, there is PDB.
I do however, think it benefits the community to shorten the Table 1, so that the critical information can be easily discerned in main text not through supplemental as most of Table 1’s are.”
ARCHIVE: Test-set-and-R-free, Twinning, Low Resolution Crystallography, PHASER, HKL2000, Parallel Expression, NCS, Missing Atoms, Trends in Crystallography, Absorption Correction, Data for Refinement and Publication, Validation.
|Copyright © NIH X-Ray Diffraction Group Maintained by Dr. Xinhua Ji|
|on the NIH-NCI-CCR-MCL server (http://mcl1.ncifcrf.gov)|