Home

The Jmol Virtual Molecular Model Kit: A Resource for Teaching and Learning Chemistry

The Jmol Virtual Molecular Model Kit:

A Resource for Teaching and Learning Chemistry


 

Otis Rothenberger, Illinois State University; Thomas A. Newton, University of Southern Maine; Robert M. Hanson, St. Olaf College; Markus Sitzmann, National Institutes of Health

Click here to download a printable  pdf version of this paper

 

Background

Until 2009, the Jmol applet was used primarily as a model display and animation tool. During the 2009 calendar year, a group of Jmol users and developers added a molecular editor to the Jmol applet capability (1). The authors of this report worked as part of that team, creating a Web-based small molecule model kit and providing enhanced applet development. The model kit served as a source of ideas as well as a testing platform for the overall project. It also provided a Web application, the Jmol Virtual Molecular Model Kit (VMK) that was released simultaneously with Jmol v.12.0.RC15 in 2010: chemagic.com/vmk or chemistry.illinoisstate.edu/osrothen/vmk.

 

VMK Overview

The VMK is designed to provide easy control panel access to Jmol’s model editing ability. This means that users can build models directly in the kit’s Jmol window. Users can add or delete atoms or groups at will. With the click of a mouse, they can convert, for example, (2R)-2-chlorobutane to its (2S) enantiomer. Changing (Z)-but-2-ene to the E stereoisomer is almost as easy. The kit’s MultiUser function allows two or more people to work with the same model at the same time, a feature that enables collaboration between teachers and students.

 

While the VMK allows users to create models “from scratch”, it also offers a rich source of starter models via links to PubChem (2) and the NIH/NCI Chemical Identifier Resolver (3,4), a Web application that translates one chemical identifier, such as an IUPAC name, into another, e.g. a SMILES, InChI, or SDF, thereby enabling the VMK to display models of, and to find information about, more than 30 million compounds! The link to PubChem provides a large library of stored SDF data, while that to Resolver generates SDF data by calculation. Using the embedded JME structure drawing applet (5,6), VMK also allows users to draw 2D structures that can be rendered as 3D models.

 

As the development of the VMK progressed, a new browser application called AKA that interfaces with PubChem and ChemSpider to construct compound-specific queries used in chemical database searches was created. Figure 1 identifies the principle data connections that VMK and AKA utilize.

 

Figure 1 

Figure 1-A flow diagram illustrating the communication pathways between the components of the VMK and AKA

 

The identifiers in these queries, SMILES, InChI, InChIKey, etc., allow rapid transfer of information from PubChem, ChemSpider, NIST, ChEBI, Wikipedia, and other data sources. AKA sends an initial query to PubChem that returns the structure, formula, and missing identifiers for the identifier that was sent in the initial query. A search algorithm then constructs several identifier-specialized queries that return links to specific chemical database information: chemagic.com/aka.

 

The Control Panels of the VMK

In addition to the Jmol display window, the VMK consists of five control panels. A brief description of the features of each control panel follows. Clicking the name of a control panel will load a YouTube video demonstrating that panel’s functionality.


 


Introduction
  • provides an introduction to VMK and identifies important contributors
  • provides access to the other control panels
  • provides a link to the Problem Manual

 

Control Panel
Molecular Editor
  • create, edit, and optimize models
  • work with multiple models
  • make structural comparisons of two models
  • find chemical identifiers such as compound names and SMILES

Control Panel
Model Tools
  • load pre-assembled models from two local databases allowing display of MEPs, MOs, and partial charges
  • link models to external searches
  • display animations and vdW surfaces
  • load templates for making models of inorganic compounds

Control Panel
2D to 3D Editor
  • convert a 2D drawing into a 3D model
  • depict 3D model as 2D JME drawing
  • compare two drawings to determine their structural relationship
  • select 2D to 3D optimization method: Jmol (UFF), PubChem (MMFF94), Resolver (CORINA)

Control Panel
Image Capture
  • copy and save an image of a line drawing of the model that is in the Jmol window
  • copy and save an image of the model that’s in the Jmol window.
  • determine R/S stereochemistry of model

 

Control Panel

  

The ability to translate 2D structures drawn in JME into fully editable 3D models makes the VMK ideally suited for classroom use. The kit’s multi-user feature allows for interaction/collaboration between teachers and students alike.

 

Additional Information

Data Bases

In addition to PubChem, VMK is connected to a local database of 2000 compounds. This local database will continue to grow. Models loaded from PubChem contain partial atomic charge data, allowing display of partial charges and MEP surfaces

 

Jmol SMILES

Jmol can create a SMILES for any displayed model. Jmol also uses an approach to SMILES comparison that eliminates the need for canonical SMILES algorithms (7).

 

Help Files

The Introduction control panel contains a link to a Problem Manual, which at the present time has 12 problems, each of which provides step-by-step instructions about a specific feature of the VMK. Each VMK control panel also has an information help link to YouTube video help.

 

References

Jmol: an open-source Java viewer for chemical structures in 3D - http://jmol.sourceforge.net/

PubChem SD File Formatted Data - ftp://ftp.ncbi.nih.gov/pubchem/specifications/pubchem_sdtags.pdf

The NIH - NCI/CADD Chemical Identifier Resolver - http://cactus.nci.nih.gov/chemical/structure

NCI/CADD Chemical Identifier Resolver: Indexing and Analysis of Available Chemistry Space, Markus Sitzmann, National Cancer Institute (NCI) - http://www.int-conf-chem-structures.org/fileadmin/user_upload/lectures/ICCS2011_D8_Sitzmann.pdf

JME Molecule Editor - http://www.molinspiration.com/jme/index.html

P. Ertl, Molecular structure input on the web, J. Cheminformatics 2010, 2:1 - http://www.jcheminf.com/content/2/1/1

A Universal SMILES String Comparator: Robert M. Hanson, St. Olaf College - http://chemapps.stolaf.edu/jmol/docs/examples-12/JmolSmilesTest.htm

 

  • Login to post comments
  • 2969 reads

Comments

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
December 2, 2011 - 4:28pm — Tom Newton

Jmol VMK Journal

We would like to call your attention to a new feature on the Jmol Virtual Molecular Model Kit's YouTube channel: The Jmol VMK Journal and the Jmol VMK Journal K-12.  Each month these journals will publish a video presenting ideas for chemistry teachers at all levels regarding the use of the VMK to enhance student learning.  Volume 1 of the Jmol VMK Journal describes the use of the VMK to illustrate the correlation between boiling point and surface area for a series of linear and branched alkanes.  

The Journal is available at http://www.youtube.com/jmolvmk.

We welcome suggestions and/or contributions for future volumes.

We have also created a Users' Guide to the VMK that we will publish shortly in ePub format.  It will be available for free to interested parties.

November 9, 2011 - 12:35pm — osrothen

Author's Comment on SMILES

One very important feature of Jmol is the inclusion of SMILES generation for any model. VMK leans heavily on this feature. It's the reason that VMK can find the name of a model, talk to databases, compare models (SMILES and SMARTS), and much more.

To my knowledge, Bob Hanson has implemented a truly unique approach to SMILES and SMILES comparison. It eliminates the need for a canonical SMILES algorithm when comparing models via SMILES strings. Without going into technical detail, Bob's approach matches a query SMILES with a given target SMILES. The script command structure for a SMILES and a SMARTS query is shown below:

target.find("SMILES", query)              target.find("SMARTS", query)

The find process is carried out in three distinct phases:

1) Create a connectivity graph of the target SMILES string.
2) Search that model for a match to the nonstereochemical SMILES query.
3) Check for stereochemistry.

In the above command structure description, I am using Bob's words. The technical details are way over my head, but I can see the bottom line of 1-3 above: SMILES and SMARTS identity matching no longer involves exact string matching. In my opinion, this little known Jmol feature represents a very significant SMILES breakthrough development.

VMK is not just a Prentice Hall model kit made up of bits and bytes. It's a model kit that can talk to the outside world. It can ask questions and it can receive answers. Jmol SMILES coupled to the Chemical Identifier Resolver is what makes this communication possible.

Best Regards, Otis

November 8, 2011 - 5:35pm — Robert E. Belford

AKA and XCITR

I am very very impressed with AKA and would like to quote your video:,
http://www.youtube.com/user/JmolVMK#p/u/8/7_f56RfBzJM
"If you are lost in the information jungle, let AKA be your guide".

Now, the first two papers of the Newsletter dealt with Chemical Information Resources and I am sort of addressing this to the authors of those papers in addition to yours. Do tools like AKA belong in the library of the future?

That leads to, are services like AKA needed to be part of the information competencies for undergraduates? Should they be in XCITR? Are they in XCITR? Of course that gets back to how sustainable and endurable your sites are, I mean does ISU/USM ensure the future of your site the way RSC does ChemSpider?.... What is the role of AKA? It seems to me it has a role outside of VMK, and VMK is sort of a portal into AKA, which in itself is a portal into ChemSpider/PubChem/?/?/....

What are your visions with AKA?

Cheers,
Bob

 

November 9, 2011 - 12:59am — osrothen

AKA

Bob,

AKA was developed as a stand alone application, but the catalyst for this involved two factors related to Jmol and VMK. First, Jmol generates a SMILES for any model in the display window. Second, this Jmol SMILES allowed us to connect VMK to the NIH/NCI Chemical Identifier Resolver. As an aside here, Bob Hanson, Jmol's project leader, has developed a truly unique approach to SMILES in Jmol, and I want to address this is a separate post later today. As indicated in our paper, Resolver became VMK's widow to the world of chemical databases. While working with Marcus Sitzmann, the creator and manager of Resolver, we came to a better understanding of the details of cheminformatics. Specifically, we came to appreciate that some of the nitty gritty details of cheminformaticts belong in the chemistry curriculum. AKA was designed to expose students to these details. AKA is an on going project. Most recently we have moved to social media in an attempt to stimulate student/teacher interest:

https://twitter.com/#!/CheMagicOrg

With respect to ensuring the future of the CheMagic domain, I have some options at ISU. In 2006, I discussed this with the College of Arts and Sciences IT administrator, and I need to follow up on the suggestions that were made at the time. I also need to find a young faculty member or graduate student caretaker for this domain.

Best Regards, Otis

November 9, 2011 - 10:29am — Robert E. Belford

Cheminformatics OLCC

Otis,

I too have come to "appreciate that some of the nitty gritty details of cheminformaticts belong in the chemistry curriculum".  Are you familiar with the OLCCs the CCCE has run in the past (see link at top of the CCCE page)? 


We wish to run one an OLCC on cheminformatics which will allow schools that do not have a cheminformatician to offer a course in cheminformatics.  A sort of "Essential Cheminformatics for Undergraduates" course.  We are seeking some funds and with some luck this will become a reality which will be available to anyone who is interested.  If you check the CCCE blogs you will see one on the OLCC, and the first entry http://www.ccce.divched.org/node/67 describes some of the issues.  One of the things we wanted to do was create TLOs (Teaching and Learning Objects) like your YouTubes and makes those available for repurposing in other courses. 

At the end of this newsletter we were going to spend a couple of days discussing several CCCE initiatives of which the OLCC is just one.

Cheers,
Bob

November 9, 2011 - 12:21pm — osrothen

OLCCs and Author Ramblings

Bob,
Thanks for pointing me to these Web pages and introducing me to OLCCs.

Since meeting Markus, I've become almost evangelical on this issue! Two simple cases in point:

1) PubChem SDF's carry with them PubChem MMFF94 partial charges. Of course, VMK imports these with each model load from PubChem. A convention used in these charges is that alkane like C-H partial charges  are simply recorded as zero - e.g. for ethanol the data is:

<PUBCHEM_MMFF94_PARTIAL_CHARGES>
3
1 -0.68
2 0.28
9 0.4

Despite the fact that there is a logical cheminformatics reason for this, I've talked to chemists who view these partial charges as useless. They are missing the point.

2) On the subject of chemical identifiers, I've talked to organic chemists who see no reason to include the details of identifiers like SMILES, InChI, and InChIKey into the curriculum. When queried on the the inclusion of IUPAC, they say, "Of course." IUPAC, InChI, and SMILES all hold something in common. They are all maps of the entire structure of a molecule. We choose to teach only the mapping system that predates the age of modern information technology!

OK, I'll stop! But first let me point out a wonderful IUPAC resolver site:

http://opsin.ch.cam.ac.uk/opsin/1,3,7-trimethylpurine-2,6-dione.png

This site is amazing, and in truth, just flat out fun!

Best Regards, Otis

November 8, 2011 - 8:54am — Robert E. Belford

Site Hosting

Otis and Tom,

This question deals with the hosting of VMK, as I am sensing that you are looking for people to adapt, use and advance classroom applications of this technology.  For someone to integrate it into their curriculum and make it a mandatory component I would think they would need some sort of guaranteed access to, and control/responsibility of the VMK's operations.  I know Jmol is open source and available through SourceForge, was VMK developed as an open-source collaborative project?  It also seems that VMK has some completely different types of features.  I notice a lot of it uses JavaScript, is any server side based?

How would you envision someone hosting the VMK?  Is this a service you would provide?  I am just trying to figure out what model you have for making this available to others in a way where people could adapt it in a sustainable manner.

Cheers,
Bob

 

November 8, 2011 - 12:28pm — osrothen

VMK as a Service

Bob,

I'm really glad you asked this question!

The Web application VMK is not Open Source at this point, but we offer it as a free service. If it proves useful to users, we will continue offering free access to this service. CheMagic is really a chemistry education service of the Chemistry Department at Illinois State University (ISU), and VMK exists on the ISU site as well as on an identical CheMagic domain. I should explain that VMK itself is a joint venture involving ISU and the University of Southern Maine. The CheMagic domain has offered free chemical education resources for over ten years.

The VMK application uses a combination of client side JavaScript, AJAX, and server side scripting (ASP.NET). The server side programming is used primarily to load data from external databases without the need of the signed Jmol applet. AJAX is used to load this remote data into Jmol without having to reload the applet. Javascript is used to access the native Jmol script modeling commands that were introduced into Jmol during the course of this project.

Over the years, Tom and I have shared a view that was held by many organic chemists going back to the days of Chime: It should be possible to put the essence of a Prentice Hall model kit into a Web application. When we retired we decided to make this happen. This led to a very successful collaboration with the Jmol developers. We are now very interested in pushing this concept further with additional collaborators. Such collaboration could well be directed toward moving the application to other servers.

Otis

November 7, 2011 - 7:04pm — kelly_lancaster

Jmol and PhET

I have one comment and one question:

We are using Jmol in a new PhET sim on molecule polarity. The model tools in the "Real Molecules" tab were very engaging for students:

http://phet.colorado.edu/en/simulation/molecule-polarity

Did you do any testing of the VMK interface with students? We have learned a lot about interface design by doing interviews.

November 8, 2011 - 8:50am — osrothen

Student Feedback

Kelly,

Thanks for pointing out the use of Jmol in the PhET module. Your query about interface testing gets to an issue where we need some help. As Tom implied in his note to Bob, we designed VMK from the ground up with two thoughts in mind:

1) Provide a Web application that could help Jmol developers work toward the inclusion of native molecular modeling script commands in Jmol.

2) Create a Web application that was better than a classic plastic student model kit.

During this process, our only feedback was from Jmol developers and other chemists. We sorely need student feedback. As Tom indicated, we are hoping that this article will stimulate some student/teacher use. Subsequent feedback from teachers and students will be most welcome.

Best Regards, Otis

November 7, 2011 - 4:03pm — Robert E. Belford

Multi User

Wow, I have been watching your YouTube Videos and this is great!  I have two questions about the multi-user interface. I went to one of your Youtubes which I don't think the paper links

http://www.youtube.com/user/JmolVMK#p/u/7/fqhnV9v10O8

First, is anyone using this in a class? Do you have an examples of lesson plans.

Second, OK, don''t get mad at me. I went in and created a compound with id 123456789.  Then asked myself what if someone else created a compound with the same ID and saved it to multiuser, so I did that with a compound and gave it an id 111222333 and destroyed your compound....sorry.  Now, I believe you state an ID is stored for 7 days, how would you create an ID for a lesson plan and be certain the molecule does not get changed? Do you keep a history of molecules associated with an ID? 

This is really great stuff and I am looking forward to playing around with AKA. 

Thanks for sharing it.

Cheers,

Bob

 

November 7, 2011 - 10:27pm — Tom Newton

Lesson plans

Bob,

Otis and I have used Jmol in our courses for many years, going back to the early days of Chime. The VMK was developed in conjunction with Jmol development over the past few years, so the kit has only recently arrived in its present incarnation.  There are a few people who are using the VMK in their courses already, mostly colleagues I know personally.  Our objective in submitting this paper is to spread the word so that others will be aware of the utility of this app and start to use it in their courses.

Both Otis and I are retired, so we have not developed lesson plans that involve the VMK.  The closest thing we have to lesson plans at the moment are the exercises in the Problem Manual.  If you look at them, they should give you a few ideas of ways to use the kit in your courses.  We continue to work on these and will add new exercises as time goes by.  We hope that others will devise new ways to use the app in their courses and share them with us and with all interested parties.  

Tom

 

 

November 7, 2011 - 8:54pm — osrothen

Multi-User Feature

Bob,

I'm going to let Tom address your usage and lesson plan questions. I'll handle the multi-user question.

Don't worry about over writing our files. Our ultimate objective here is to give intereted users a dedicated account. As it stands now, a user must pick a reasonably obscure 9 digit number. The selected number then corresponds to the name of a text file that holds the Jmol state of the model being saved. This file is overwritten each time a user saves a model. The 9 digit number is held in a local cookie for 7 days. When this time elapses, the file in not erased. The user simply needs to log on again.

Accidental overwriting caused by groups selecting identical numbers by coincidece is a problem. We will address this issue with dedicated accounts if this feature becomes popular.

Best Regards, Otis

November 7, 2011 - 2:05pm — malkayayon

Virtual Molecular Model Kit

The Virtual Molecular Model Kit is intuitive and easy to use.  will use it and show other teachers so they use it too.


I can only think of positive comments. The only problem I saw was the possibility to add more bonds than those that are possible.


The tutorials are great!!


Great job!


Malka

November 7, 2011 - 4:59pm — Tom Newton

extra bonds

Malka,

Thanks for your feedback.  A few words about making models.  If you have the Atom:O option selected and you click on an H atom of ethane, for example, it generates ethanol, i.e. it replaces the H with an OH.  If  you click on one of the two C atoms in ethane, you get methanol, i.e. it replaces the CH3 group with an OH group. If you click on a C and drag into the window, you get a model with the formula C2H7O (which appears in the calculator window at the top of the page), i.e. this action makes a C-OH bond.  The correctH function will return ethanol, C2H6O.  The tutorial on using templates illustrates the different results obtained by clicking vs. click-dragging.

 

Tom

November 7, 2011 - 3:22pm — osrothen

Number of Bonds

Malka,

Thank you for the kind words. One of the objectives for presenting this paper is to make contact with teachers who may use this tool. If you do find it useful, please keep in touch with us regarding suggestions from students and teachers.

The ability to create more bonds to an atom than allowed (except when H atoms are present) is actually useful - e.g. showing transition states.

We do have an auto rectification of H atoms feature, and I must admit there are occasional problems with this feature. Jmol uses geometry to set H atom rectification values, and sometimes a model edit will leave the model in a geometry that does not rectify H correctly. When this happens, several features can be used to hand edit the model - Atom: drag; Atom: delete; and Model: correctH. Again, Model: correctH can fail depending on atom geometry after the edit. When this happens, one of the other methods must be used.

Otis

November 8, 2011 - 10:30am — malkayayon

translating the tutorials

Hi,


Thanks for your explanations. If I want to use the tutorials in Israel it would be good to translated them.


Is it possible to translate your tutorials to hebrew, and credit you?


Thanks


Malka

November 8, 2011 - 5:21pm — Tom Newton

translation

Hi Malka,

When you say tutorials, are you referring to the exercises in the Problem Manual?  Neither Otis nor I have a problem with that, but we'd appreciate it if you would clarify what you want to translate.

 

Thanks.

 

Tom