Discussion:
[BlueObelisk-discuss] Two Google Summer of Code projects related to the CDK
Egon Willighagen
2016-03-01 13:49:43 UTC
Permalink
Hi all,

the NRNB is an USA-based organization around Cytoscape and
WikiPathways, the latter involving PathVisio. You may know that
Cytoscape has a plugin that used the CDK (by Scooter Morris et al.,
UCSF) and here in Maastricht we have been working on PathVisio plugins
for metabolites also using the CDK.

Now, the NRNB has been accepted to the Google Summer of Code and there
are at least two projects where the CDK is related too:

1. https://github.com/nrnb/GoogleSummerOfCode/issues/46
The following project idea is using Java and would involve the CDK directly:

https://github.com/nrnb/GoogleSummerOfCode/issues/46

Here, the idea is integrating a 2D visualization plugin and the
outcome of work by Rianne Fijten for metabolite data (e.g. using the
CDK-based HOSE/NMR prediction stack). These plugins need updating and
integration.

2. https://github.com/nrnb/GoogleSummerOfCode/issues/47
The second idea is in JavaScript where I envision use of remote
services that can convert SMILES and/or identifiers to SVG images of
2D depictions. John's CDKDepict would be a good candidate, but I would
not mind a flexible system that can use other services too, like
perhaps AMBIT.

Now, Noel asked about a JS translation of the CDK stack. That would be
a lot more work, and not sure if it is entirely feasible, but I'm more
than happy to discuss that with the student applicant and co-mentor.

So, if you are a student or know one, please check out these ideas. Of
course, applying does not mean you will get the 5000 dollar to work
for three months during the summer holidays. That's because the NRNB
organization will pick the best proposal (which the students will have
to write). The project ideas are just suggestions, and students are
invited to tune to their interest (though commonly in collaboration
with the anticipated mentor(s)).

Egon
--
E.L. Willighagen
Department of Bioinformatics - BiGCaT
Maastricht University (http://www.bigcat.unimaas.nl/)
Homepage: http://egonw.github.com/
LinkedIn: http://se.linkedin.com/in/egonw
Blog: http://chem-bla-ics.blogspot.com/
PubList: http://www.citeulike.org/user/egonw/tag/papers
ORCID: 0000-0001-7542-0286
ImpactStory: https://impactstory.org/EgonWillighagen
Geoffrey Hutchison
2016-03-01 15:55:22 UTC
Permalink
Post by Egon Willighagen
2. https://github.com/nrnb/GoogleSummerOfCode/issues/47
The second idea is in JavaScript where I envision use of remote
services that can convert SMILES and/or identifiers to SVG images of
2D depictions.
There may be benefits to using CDK for this, but I think there are already ways for JS to generate SVG from SMILES:

Helium:
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-helium.html

RDKit:
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-rdkit.html

Open Babel:
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-open-babel.html

Cheers,
-Geoff
Egon Willighagen
2016-03-01 17:51:34 UTC
Permalink
Geoff,

thanks for the suggestions. I need to look at the HTML code that uses
$foo.js and create an SVG image... emscripting looks nice, but also
harder to easily integrate... I need to see how much you can rip out,
to just have the core content...

Egon
Post by Geoffrey Hutchison
Post by Egon Willighagen
2. https://github.com/nrnb/GoogleSummerOfCode/issues/47
The second idea is in JavaScript where I envision use of remote
services that can convert SMILES and/or identifiers to SVG images of
2D depictions.
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-helium.html
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-rdkit.html
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-open-babel.html
Cheers,
-Geoff
--
E.L. Willighagen
Department of Bioinformatics - BiGCaT
Maastricht University (http://www.bigcat.unimaas.nl/)
Homepage: http://egonw.github.com/
LinkedIn: http://se.linkedin.com/in/egonw
Blog: http://chem-bla-ics.blogspot.com/
PubList: http://www.citeulike.org/user/egonw/tag/papers
ORCID: 0000-0001-7542-0286
ImpactStory: https://impactstory.org/EgonWillighagen
Greg Landrum
2016-03-01 18:06:16 UTC
Permalink
I can probably help with that Egon, at least as far as the RDKit bit is
concerned.
Post by Egon Willighagen
Geoff,
thanks for the suggestions. I need to look at the HTML code that uses
$foo.js and create an SVG image... emscripting looks nice, but also
harder to easily integrate... I need to see how much you can rip out,
to just have the core content...
Egon
Post by Geoffrey Hutchison
Post by Egon Willighagen
2. https://github.com/nrnb/GoogleSummerOfCode/issues/47
The second idea is in JavaScript where I envision use of remote
services that can convert SMILES and/or identifiers to SVG images of
2D depictions.
There may be benefits to using CDK for this, but I think there are
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-helium.html
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-rdkit.html
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-open-babel.html
Cheers,
-Geoff
--
E.L. Willighagen
Department of Bioinformatics - BiGCaT
Maastricht University (http://www.bigcat.unimaas.nl/)
Homepage: http://egonw.github.com/
LinkedIn: http://se.linkedin.com/in/egonw
Blog: http://chem-bla-ics.blogspot.com/
PubList: http://www.citeulike.org/user/egonw/tag/papers
ORCID: 0000-0001-7542-0286
ImpactStory: https://impactstory.org/EgonWillighagen
------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
_______________________________________________
Blueobelisk-discuss mailing list
https://lists.sourceforge.net/lists/listinfo/blueobelisk-discuss
Egon Willighagen
2016-03-01 18:11:38 UTC
Permalink
My first go will probably to have SVG remotely generated... John's
CDKDepict can do this, and so does Nina's AMBIT... that reduces the
dependencies for pvjs significantly...

But I'll keep it in the picture. BTW, please feel free to add these
suggestions to the GitHub NRNB issue tracker too, so that students can
see the ideas too!

Egon
Post by Greg Landrum
I can probably help with that Egon, at least as far as the RDKit bit is
concerned.
Post by Egon Willighagen
Geoff,
thanks for the suggestions. I need to look at the HTML code that uses
$foo.js and create an SVG image... emscripting looks nice, but also
harder to easily integrate... I need to see how much you can rip out,
to just have the core content...
Egon
Post by Geoffrey Hutchison
Post by Egon Willighagen
2. https://github.com/nrnb/GoogleSummerOfCode/issues/47
The second idea is in JavaScript where I envision use of remote
services that can convert SMILES and/or identifiers to SVG images of
2D depictions.
There may be benefits to using CDK for this, but I think there are
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-helium.html
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-rdkit.html
http://baoilleach.blogspot.com/2015/02/cheminformaticsjs-open-babel.html
Cheers,
-Geoff
--
E.L. Willighagen
Department of Bioinformatics - BiGCaT
Maastricht University (http://www.bigcat.unimaas.nl/)
Homepage: http://egonw.github.com/
LinkedIn: http://se.linkedin.com/in/egonw
Blog: http://chem-bla-ics.blogspot.com/
PubList: http://www.citeulike.org/user/egonw/tag/papers
ORCID: 0000-0001-7542-0286
ImpactStory: https://impactstory.org/EgonWillighagen
------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
_______________________________________________
Blueobelisk-discuss mailing list
https://lists.sourceforge.net/lists/listinfo/blueobelisk-discuss
--
E.L. Willighagen
Department of Bioinformatics - BiGCaT
Maastricht University (http://www.bigcat.unimaas.nl/)
Homepage: http://egonw.github.com/
LinkedIn: http://se.linkedin.com/in/egonw
Blog: http://chem-bla-ics.blogspot.com/
PubList: http://www.citeulike.org/user/egonw/tag/papers
ORCID: 0000-0001-7542-0286
ImpactStory: https://impactstory.org/EgonWillighagen
Robert Hanson
2016-03-01 18:11:37 UTC
Permalink
Post by Egon Willighagen
Hi all,
Now, Noel asked about a JS translation of the CDK stack. That would be
a lot more work, and not sure if it is entirely feasible, but I'm more
than happy to discuss that with the student applicant and co-mentor.
Now THERE's a challenge! Java2Script could probably manage this. But in my
experience, performance would suffer considerably without careful attention
to detail. The cool thing, of course, is that if you used Java2Script, then
you would have a consistent Java/JavaScript code synchronization. This is
probably the
coolest thing about Java2Script. You get to keep going with all development
in Java only. Not a one-time port.

Note that you would need full source for all accessed third-party JAR files.

A quick look at the CDK suggests openscience/cdk has something like 15,000
methods. The initial conversion would be relatively easy, provided all the
source there; validation and optimization would be the real chore.

I only have experience with Java2Script. I love it. Anyone have any other
ideas?

BTW I am not volunteering to do this....

Bob

Bob

Loading...