Imperial College London

Henry Rzepa

At the ACS conference, I have attended many talks these last four days,
but one made some "connections" which intrigued me. I tell its story (or
a part of it) here.

But to start, try the following experiment.

1.  Find a Word document of .docx type on your hard drive
2.  Remove the .docx suffix and replace it with a .zip suffix.
3.  Expand as if it is an archive (it is!).
4.  A folder is created and this itself contains four further folders.
    These all contain XML files, and in the sub-folder actually called
    word you will find something called **document.xml** That file
    contains the visible content of the document; all the others are
    support documents, including styles etc.

The reason this is important was made clear in Santi Dominguez' talk.
Most of it was concerned with introducing Mbook, an ELN (electronic
laboratory notebook) but the relevance to the above comes from his
introduction of **Mpublish**, a forthcoming product targeting the area
of research data management. What is the connection? Well, NMR
spectrometers produce raw outputs as collections of files, much in the
manner of the exploded word document above. Some files contain the raw
FID, others contain the acquisition parameters, etc. These files are
then turned into the traditional spectra by suitable processing software
such as Mestrenova (part of the same ecosystem as Mpublish). Most users
of such programs then squirt the spectra into a PDF file and it is this
last document that is preserved as "research data" -- almost invariably
this is the version sent off to journals as the supporting information
or SI for the article. SI is called information for a good reason; in
such a container it is very often not easily usable data, and functions
just visually.

So what **is** the problem? Well, the conversion of the NMR fileset (and
quite possibly many other forms of spectroscopy) into a PDF file is a
lossy process. It cannot be reversed; information has been lost. And
only really a human who can easily retrieve and interpret such a visual
presentation.

Santi described how Mpublish can assemble all the files associated with
the instrumental outputs, optionally add chemical structure and other
information, collect suitable metadata describing the contents and
create a .zip archive. As we saw with Word however, the suffix does not
even need to be .zip. It was suggested that it be this
information-complete archive that should really be used as SI to
accompany an article in which NMR data is invoked to support the
narrative. In the reverse process, anyone downloading this zip archive
could themselves potentially acquire full access, without information
loss, to the original NMR data. There is a little further magic that
needs to be included to make the process work which I do not include
here. When Mpublish becomes available to play with, I will complete that
story here.

It is good to report that software is starting to appear which enhances
the management and reporting of research data as part of the publication
process. The "rules" and "best practice" of this game are still being
written however. In this regard, I feel that it is the researchers
themselves that must play a vital role in defining the rules. Let us not
cede that role just to publishers.

::: {.wp_orcid_field}
[https://orcid.org/0000-0002-8635-8390](https://orcid.org/0000-0002-8635-8390){target="_blank"
rel="author"}
:::


At the ACS conference, I have attended many talks these last four days, but one made some “connections” which intrigued me. I tell its story (or a part of it) here. But to start, try the following experiment. Find a Word document of .docx type on your hard drive Remove the .docx suffix and replace it with a .zip suffix. Expand as if it is an archive (it is!). A folder is created and this itself contains four further folders.

Messaggi di Rogue Scholar

Research data: Managing spectroscopy-NMR.