[Rp] Structural flexibility in proteins: impact of the crystal environment

[khinsen]

Original article: Hinsen, K (2008) Structural flexibility in proteins: impact of the crystal environment. Bioinformatics 24:521. (Open Access version)

PDF URL: https://github.com/khinsen/rescience-ten-year-challenge-paper-3/raw/master/article/article.pdf
Metadata URL: https://github.com/khinsen/rescience-ten-year-challenge-paper-3/raw/master/article/metadata.yaml
Code URL: https://github.com/khinsen/rescience-ten-year-challenge-paper-3/

Scientific domain: Biophysics
Programming language: Python
Suggested editor: @delsuc

[rougier]

@delsuc Can you edit this submission to the "Ten year reproducibility challenge" ?

[delsuc]

ok, no problem.

[rougier]

Thanks. Note that you need only one reviewer (because this is a reproduction)

[delsuc]

@pdebuyl could you review this submission ?

[pdebuyl]

Hi @delsuc

I reviewed the first submission. I think that it would be best to get another reviewer to proof the process for the ten year challenge. I am ok to review anyway after next week if necessary.

[delsuc]

@fxcoudert would you be able to handle this review ?

[fxcoudert]

@delsuc Here is my review

Paper

I quite appreciated the non-nonsense writing of the paper, which is good for a reproducibility study to show the pitfalls and issues, even when they turn out not to be crucial.

• Regarding the numpy evolutions, which I think are quite important to the community at large, maybe this could be better characterised: what is the numpy policy on breaking changes, deprecation process, and use of semantic versioning? Maybe links to this information could be added. (But that opens another 10-year challenge: if a paper has web links, how many will still be active 10 years later?)
• The instructions for reproduction are rather clear, and the recommendation to use Guix is interesting (I know of it, and want to try it myself but haven't found the time yet).
• The main point I think could be discussed is the use of Python 2, given the EOL was earlier this year (2020-01-01), and it is being removed from software distributions. How long does the author suggest it will be feasible to install Python 2 on modern hardware? Or without a dedicated OS like Guix? I would be interested in comments on that, as I think it's a worrying topic.
• Grace is still working fine, I am using it daily :)
• I do not understand why there are two folders named updated-scripts-from-obsolete-web-server and scripts-from-suplementary-material-to-original-paper, and they are not documented in the paper. One of them has an extra script (calculate_crystal_fluctuations_small_memory.py) which is not mentioned anywhere else either. Edit: I see that the notes file mentions those. Probably best to mention that in the paper itself, though.

Typo

• “there is longer any point in such a change” (page 1)

[delsuc]

Thank you very much @fxcoudert !
@khinsen can you improve the document as mentionned ?