Saturday, January 25, 2020

MetaboEU2020 in Toulouse and the ELIXIR Metabolomics Community assemblies

This week I attended the European RFMF Metabomeeting 2020, aka #MetaboEU2020, held in Toulouse. Originally, I had hoped to do this by train, but that turned out unfeasible. Co-located with this meeting where ELIXIR Metabolomics Community meetings. We're involved in two implementation studies for together less than a month of work. But both this community and the conference are great places to talk about WikiPathways, BridgeDb (our website is still disconnected from the internet), and cheminformatics.

Toulouse was generally great. It comes with its big city issues, like fairly expensive hotels, and a very frequent public transport system. It also had a great food market where we had our "gala dinner". Toulouse is also home to Airbus, so it was hard to miss the Beluga:

The MetaboEU2020 conference itself had some 400 participants, of course, with a lot of wet lab metabolomics. As a chemist, with a good pile of training in analytical chemistry, it's great to see the progress. From a data analysis perspective, the community has a long way to come. We're still talking about known known, unknown knowns, and unknown unknowns. The posters were often cryptic, e.g. stating they found 35 interesting metabolites, without actually listing them. The talks were also really interesting.

Now, if you read this, there is a good chance you were not at the meeting. You can check the above linked hashtag for coverage on Twitter, but we can do better. I loved Lanyrd, but their business model was not scalable and the service no longer exists. But Scholia (see doi:10.3897/rio.5.e35820) could fill the gap (it uses the Wikidata RDF and SPARQL queries). I followed Finn's steps and created a page for the meeting and started associated speakers (I've done this in the past for other meetings too):

Finn also created proceedings pages in the past, which I also followed. So, I asked people on Twitter to post their slidedeck and posters on Figshare or Zenodo, and so far we ended up with 10 "proceedings" (thanks to everyone who did!!!):

As you can see, there is an RSS feed which you can follow (e.g. with Feedly) to get updates if more materials appears online! I wish all conferences did this!

Thursday, January 16, 2020

Help! Digital Object Identifiers: Usability reduced if given at the bottom of the page

The (for J. Cheminform.) new SpringerNature article template has the Digital Object Identifier (DOI) at the bottom of the article page. So, every time I want to use the DOI I have to scroll all the way down to the page. That could be find for abstracts, but totally unusable for Open Access articles.

So, after our J. Cheminform. editors telcon this Monday, I started a Twitter poll:

Where I want the DOI? At the top, with the other metadata:
Recent article in the Journal of Cheminformatics.
If you agree, please vote. With enough votes, we can engage with upper SpringerNature manager to have journals choose where they want the DOI to be shown.

(Of course, the DOI as semantic data in the HTML is also important, but there is quite good annotation of that in the HTML <head>. Link out to RDF about the article, is still missing, I think.)

Sunday, December 29, 2019

Groovy Cheminformatics for CDK 2.3

Last week I worked on my Groovy Cheminformatics book and made a first release for CDK 2.3 (doi:10.5281/zenodo.3590374). One of the new features is that the Groovy code runs on the command line without having to download or install something first:

The CreateAtom3.groovy example with @Grab instruction.
There is yet plenty to migrate from the older paper version and, for example, the Depiction chapter is still missing.

Anyways, I hope you like it :)

WikidataCon, Google Summer of Code Mentor Summit, OpenRiskNet, and the Beilstein Open Science meeting

OpenRiskNet workshop material.
Autumn has been hard. Due to an unfortunate combination of events, the workload was hardly bearable. Half December it started to slow down, but not enough. That said, I had a pleasant two weeks in October, when I attended four meetings in two weeks time. Busy as well, but inspiring. Normally, I blog about meetings I attend, but I haven't had time for that yet. In fact, that has become more common in the past few years.

Beilstein Open Science
The first in the row was the Beilstein Open Science Symposium. It was the second time for me to attend this meeting. It had a nice line up, but would actually not mind to have a hackathon component too. I spoke at this meeting about Wikidata and Scholia, of course (my slides are available at doi:10.5281/zenodo.3492008). It was good to see updates of various projects around, see this abstract book.

Google Summer of Code Mentor Summit
From the meeting I walked down from the venue down to the train station through the grape vines fields. Very enjoyable. The train to Frankfurt Hbf takes long, and the train to Muenchen was delayed by an hour. So, I missed the formal opening of the Mentor Summit, but in time for the goodies. The meeting was awesome: 15 years of open source projects, with plenty of science projects. I was there as mentor of the NRNB organization, but I proudly wore my first GSoC t-shirt from 2007, impressing even the local organization :) Back then, Alexander worked on chemistry search support for the KDE desktop.

OpenRiskNet Final Workshop
Next up was the OpenRiskNet Final Workshop, held in Amsterdam. Just far enough from home to need a hotel. Because of overlap with the next meeting, I even had to skip the last afternoon session. Running from one meeting to another, I actually had just one day to prepare, which was tight, as I was going to give a workshop. Fortunately, I had plenty of material, and it just came done to putting it in the right order. I decided to try something new, and use Markdown (git: OpenRiskNet/workshop/OntologyWorkshop). I copied the idea of the prev/next buttons from one of the other workshops in the same git repository, but am particularly happy with the Answer button, which toggles the answer (see above screenshot).

The final meeting in this series was WikidataCon (or this Scholia page)in Berlin. My abstract around Wikidata and Scholia was accepted and I joined in the informal hack sessions. The latter resulted in an externally developed Scholia feature (see doi:10.3897/rio.5.e35820), so cannot complain. In case you missed it, Wikidata is booming. In fact, the bigger problem with Wikidata, right now, are its growing pains, repeatedly visible in systems not being able to keep up with the effort. Beside my blog, I can recommend this preprint, if you want to read up on Wikidata. My talk, Cheminformatics to improve Wikidata on chemical compounds,  was actually recorded, and you can watch the video here.