Monday, February 28, 2011

Eduroam now available to UC chemistry staff

Students have long had better access to wireless internet than staff in the Department of Chemistry.  But this changed when we got our new and hard to remember user names.  It's just that no-one (I am looking at you, Science-IT) bothered to tell us.   Having said that, when I asked Science-IT I got the following helpful tip:

Follow the instructions for students on this site, and substitute <ki-id> for <ki-id>

Friday, February 25, 2011

Web applications in KemiKS: Quiz and Hückel

This week, we have tried three new visual features in KemiKS.

a) The students were strongly encouraged to use Maple in the solution of one problem, where they are supposed to draw the wavefunction of the hydrogen molekyle ion. This exercise was previously solved by hand, but now was supposed to be solved by using Maple. The students responded positively to it, although they had lots of problems in remembering how to use Maple. But when they finally had plotted the wavefunction, they really liked the fact that they now had created themselves the figures shown in many textbooks.

b) Janus had found a website at, with a good discussion of atomic orbitals up to f-orbitals and a nice quiz. The student were asked to do the quiz in the first exercise class on monday as a repetition of what they had learned in the week on atomic orbitals. We will later in week 7 follow up on this idea and give the students in the beginning of the week a quiz on the key points of the week, which they probably will not be able to answer. At the end of the week, the students will then repeat the quiz and hopefully can see what they have learned in this week. Janus will look for appropriate tools in order to make a quiz with high score and solutions outside of Absalon.

c) as illustration of the discussion of the Hückel method, we have now incorporated two java based Hückel programs (see description in previous blog), which can be run over the net or can be downloaded and run directly on ones computer. This was thought to have the big advantage that we can study much larger systems, which previously were out of question, when one had to diagonlize the Hückel matrices by hand. The students were then asked to use these tools in two of the problems of this week. One problem was newly made and the other was modified so that the students now use these Hückel programs instead of doing it by hand. Talking to the teaching assistants helping the students the response is mixed. On one hand the Hückel programs are nice tools, but the danger exist that the students loose the understanding of how the Hückel matrix should look like, because in these Hückel programs it is generated automatically for them after drawing the molecule in the GUI. Therefore, it might actually didactically more effective to let the students write down the Hückel matrix by hand and then use Maple to diagonalize it, which is without computers typically the limiting factor or to have to exercises: one with the Hückel matrix by hand / Maple approach for a medium size molecule and one with the Hückel programs for a large molecule.

Sunday, February 20, 2011

Smelling vibrations: why do isotopomers have different smells?

An interesting paper (I assume, I haven't read it yet) led to an interesting blog post (I did read that one) with an interesting discussion (be sure to check out the comments below, including those of the paper's author). 

Science in action and science in the open.

Saturday, February 19, 2011

Symmetry: Ancient Knowledge Rediscovered

There are things in this world you never learn. You either wish you never will, or thought you never would. When the topic concerns symmetry of molecules, I've used (and keep using) the good old C1 for quite some time without any hassle. Excuses for not exploring it have been many, but when I had to be assistant teacher in a course (KemiKS) which the Department of Chemistry at the University of Copenhagen offers, I had to sit down and buckle up. Visualising it in 3D proved to be the hard task, so I went on the web to find the symmetry@otterbein website. It has interactive pages where you can rotate around the molecules to get a good feel for where axis and planes are. It also includes a challenge so you can learn to use the flow charts that follows every book on the subject.

Now symmetrize a molecule!

COMS Seminar Calendar

A COMS Seminar Calendar has been created. You can add it to your favorite calendar program by using the link. This will be maintained by Casper Steinmann.

For detailed information, look at the appropriate page about the COMS Seminars on the COMS homepage.

Thursday, February 17, 2011

Center of Excellence proposal moves on to round 2

The Center of Excellence pre-proposal I, Thomas Hamelryck, Flemming Poulsen, and Jesper Ferkinghoff-Borg submitted to the Danish National Research Foundation (Danmarks Grundforskningsfond) was one of 27 selected out of 198 for the second round.

The proposal builds on Anders Christensen's Ph.D. project on protein structure determination using chemical shifts.

The proposal deadline is the 28th of April, and we'll know if we made it some time in October.

Positions in Irvine and Singapore

I have several openings for postdoctoral positions in different areas of theoretical and computational chemistry and physics in my group at UC Irvine. I will appreciate it if you could bring this to the attention of suitable candidates.
Thanks for all your help.
Shaul Mukamel
(A) Computational Biophysics
Simulation of amyloid fibril structures and aggregation kinetics, lipid-protein complexes and antibody interactions and how they can be probed by novel multidimensional spectroscopy ranging from the infrared to the ultraviolet.
(B) Attosecond X-ray Spectroscopy of Molecules
Developing time-dependent many-body approaches to nonlinear x-ray core-electron spectra and their description in terms of real-space and real-time wavepackets of electrons and nuclei. Computational tools will be implemented for the design and analysis of measurements involving multiple ultrafast optical and x-ray pulses.
(C) Energy and charge separation in photosynthetic complexes studied by nonlinear spectroscopy
Developing and applying time-dependent density functional, nonequilibrium Green?s Function techniques and exciton models for computing electronic excitations of molecular assemblies, light harvesting complexes and current-carrying molecules. Connection is made to quantum information processing and manipulation.
(D) Nonlinear Spectroscopy with Quantum Optical Fields
New optical signals which use entangled photons, pulse shaping, and coherent control algorithms are designed and simulated for probing molecular relaxations and exciton transport.
Ph.D. is required. Salary will commensurate with experience. Send a curriculum vitae, publication list and arrange for three letters of recommendation to be sent to:
Professor Shaul Mukamel
Department of Chemistry
1102 Natural Sciences
University of California, Irvine
Irvine, CA 92697-2025
949/824-7600 (phone) 949/824-4759 (fax) (website)
A 2 year postdoctoral position is available in singapore, starting from april-june 2011. The candidate should have experience of computational methods in studying protein-ligand interactions and experience with free energy methods will be an advantage. The project is aimed at fragment screening, small molecule screening, peptide design against a protein of therapeutic interest. We have recently designed and patented a peptide against this target for oncology and it is hoped that this will inspire further developments in this project. The work is in close collaboration with experimental labs (biophysical/cell & molecular biology & zebrafish/mouse models) and with the oncology division of the local hospital, with a rapid turnaround time that helps guide design. It is hoped that it will result in a molecule that will be be taken over by a small local biotech after 2 years. For further details please send cv and get 3 letters of reference sent to


Tuesday, February 15, 2011

QM/MM course: notes from week 2

QM/MM Yang Paper
View more presentations from molmodbasics.

I presented the method used in the paper by Parks, Hu, Rudolph, and Yang.
Links to the papers can be found here.

The QM region is defined on slide 2 and is treated at the B3LYP/6-31G(d) level if theory while the rest of the protein and solvent is treated by the Charmm22 and TIP3P force field, respectively.  Gaussian is used for the QM calculations and Sigma is used for the rest.

The covalent boundary is treated by Yang's pseudobond method, where the link atom is described by a pseudopotential that gives the atom a valence of 1, but an equilibrium bond length close to that of a C-C single bond.

An explicit solvation model is used, with a 12-Å shell of frozen water to handle the water-vacuum boundary. No other constraints are imposed.

An approximate TS  is located with the Quadratic String Method (QSM) - a method used to interpolate between reactants and products.  It is not clear from the paper whether the approximate TS is refined further with QST3.

Protein/solvent dynamic is handled with Yang's Minimum Free Energy Path method, where the "QM atoms are optimized in the environment of the fluctuating MM subsystem.":
1. A QM/MM energy and gradient is computed for a reference QM and MM geometry along the QSM path.
2. The charges of the QM atoms are computed, and are used to run all and all MM MD (keeping the QM geometry frozen).
3. The energy and gradients are corrected using a Boltzmann weighting scheme and approximating the QM/MM electrostatic energy as point charge interactions.
4. The gradient is then used to minimize the QM region (updating the Boltzmann correction at each step), using the trajectory geometries from 1.  This is apparently done with Gaussian using the "external" keyword (slide 10)
5. Steps 1-4 are repeated so self consistency for each point along the QSM path.
6. Upon self-consistency, the vibrational frequencies of the QM region are computed.

It is not clear from the paper what charge scheme (slide 13) the study uses.

From our discussion
This is perhaps the most sophisticated treatment of dynamics in QM/MM to date, but also the potentially most complicated and error-prone.

From a "dynamics perspective" it is best to keep the QM region as small as possible, since its conformational space is not explored extensively, but the accuracy of the QM/MM interaction energy may suffer.

The parameters in the pseudobond method is basis set dependent (and optimized for 6-31G*).  It is not clear how transferable the parameters are to larger basis sets in the QM/MM region.  One could also mix the basis and use 6-31G* near the pseudobond, but this might cause problems for short side chains such as cysteine (slide 2)

European Summerschool in Quantum Chemistry 2011

European Summerschool in Quantum Chemistry 2011

to be held at Hotel Torre Normanna near Palermo, Sicily, Italy
September 18 - October 1 2011

The European summerschool in quantum chemistry will take place for the 13th time from September 18 to October 1st, 2011. The goal of the school is to introduce the participants to modern methods and computational techniques in quantum chemistry. More and more scientists in different fields of chemistry
are using theoretical methods in their research, either as a technique per se, or as a complement to experimental work. As a result there is a growing need to acquire the background knowledge necessary for skilful use of these methods. The European Summer School in Quantum Chemistry (ESQC) was started with the aim of disseminating such knowledge. The first school was held in August 1989 under the direction of the late Bjørn Roos. The emphasis of the school is more on understanding than on the technical aspects of methods and much time will be devoted to the discussion of different electronic structure problems and the choice of appropriate methods for their solution. The course will consist of both lectures and tutorials.

Two Java-based Hückel Theory programs

I here share two examples of online Hückel approximation programs. The first is located here and denoted a) in the following, while the second example, b), may be found here. Please note, that b) may be slightly unstable towards refreshing the site but may still be adequate from a teaching point-of-view, though, as it is readily downloaded as a Java-file. Update: Program a) is available in a newer implementation here (thanks to Rafael R. Pappalardo for pointing this out in the Comments).


The first program is indeed very intuitive in the sense that a molecule is easily sketched on the designated canvas and the calculation carried out by clicking just a single button on the right side of the screen. The individual carbons are marked on the canvas by initially clicking the 'Add' button to the left and subsequently placing the atoms by clicking the mouse at the appropriate locations. The bonds (only one type of bond exists, more on this below) are made by dragging a line between two adjacent atoms, equivalent to GView, Avogadro, Jmol, etc. When the molecule has been build and 'optimized' by clicking the 'Normalize' button in the left column, the calculation is made by clicking the 'Show Data Tbl' tab at the upper right corner. The results of the calculation now appears in a seperate window, divided into:
  • MO number
  • Orbital energies
  • Occupancies of the individual MOs
  • Population tables
  • Bond tables w/ appropriate Hückel bond orders
This program furthermore enables the option of illustrating the MOs. A MO diagram is build 'on-the-fly' as the user builds the bonds between the individual atoms, and by clicking the 'Show Molecule' tab in the lower right corner, the MOs are shown in accordance with the MO diagram. Note here, that the orbital energies are shown in the lower part of the screen when the user shifts between the individual MOs by clicking the 'Up' and 'Down' tabs next to the 'Show Molecule' tab.
As touched upon in the above, this program only offers sp2-hybridized orbitals in accordance with regular Hückel theory. The bonds may be altered, though, in an indirect manner, as the alpha- and beta-parameters are modifiable.


As this program has shown to be rather unstable towards running directly from the browser, I would recommend downloading it by clicking the download link below the applet. The file is just a regular .jar file but advantages with respect to the browser applet in the fact that it is faster and more reliable. Furthermore, I will not comment on the Lewis part of the program, only the Hückel part.
In this program, the molecule is build following the same guidelines as for program a). Press 'Build' and draw on the canvas. As bonds are generated automatically in this program, the MO diagram to the right is dynamic as is the total energy, shown in the left column. By pressing the 'Results' button to the left, a seperate window appears, including:

  • The Hückel Hamiltonian (in contrast to program a)). Note that the form of the matrix is in accordance with the numbering within the molecule and may thus not resemble the expected.
  • The orbital energies in terms of the alpha- and beta-parameters.
  • The electron density matrix.
  • The atomic charges as well as the total charge.
As in program a), the degree of hybridization may be altered but in a fancier manner. By clicking the box 'Parameters' in the left side column, and subsequently clicking on one of the individual carbon atoms, the program lets the user change the atom type and hence the final Hamiltonian, etc.
UPDATE: As pointed out by Markus in the comments, you can also display MOs in HuLiS by clicking on specific MO's on the energy graph. Thanks to Markus for this comment.


- Benzene: Use this link (see below) to illustrate the benzene MOs as a supplement to the MOs from program a).
Edit. As Jan mentions in the comments, this section of chemtube3d also serves as a great tool of visualization for the MOs of benzene.

- Naphthalene vs. Azulene. Calculate the differences in the HOMO-LUMO gap. Why does these differences exist and why is azulene dark blue, while naphthalene is colorless?

- Why is it impossible to design and describe 1-methylnaphthalene within the ordinary Hückel approximation? What is the 'Hückel equivalent' to 1-methylnaphthalene?

- Topic of discussion: If we really were to attempt a description of 1-methylnaphthalene, how could the program be manipulated (Hint: Think in terms of alternative ways of describing hybridization).

Friday, February 11, 2011

Modeling Chemical Reactions (in Enzyme Active Sites)

Here are the slides from my talk "Modeling Chemical Reactions (in Enzyme Active Sites)" that I'll give on Monday (February 14) as part of the Ph.D. course Biostructure and Molecular Modeling in Drug Research over at the Faculty of Pharmacy.

Thursday, February 10, 2011

Inspiration for Education at its Best (Den gode uddannelse)

COMS received at grant from the Education at its Best (Den gode uddannelse or DGU) initiative last week (see announcement here) to hire students to help incorporate visualization and molecular simulations into chemistry education.

The best place to start it to create a page of resources for each course that lists what is currently available, much like this one.  I have created a new web page to collect these resource pages.

Here are some other links to posts from my blog: that describe how to incorporate visualization and molecular simulations into chemistry education.

Simulations in teaching physical chemistry: thermodynamics and statistical mechanics
An Atkins Diet of Molecular Workbench 
One, Two, Three, MD 
Tunneling and STM (a first stab at using Molecular Workbench to teach quantum mechanics)

Jean-Claude Bradley's links to e-resources for organic chemistry

Tuesday, February 8, 2011

Drug Design and pH

Here are the slides from my talk "Drug Design and pH" that I'll give in Thursday (February 10) as part of the Ph.D. course Biostructure and Molecular Modelign in Drug Research over at the Faculty of Pharmacy.

Friday, February 4, 2011

Shooting star in Chemistry to develop lighting-fast protein analysis

From the Departmental newsletter written by Jes Andersen:

Proteins are a bit like keys. By observing the shape of the key you can figure out which lock it will open and observing the shape of a protein allows you to reveal which functions it performs. Today's technique requires the efforts of a small research team for several months to discover the configuration of a just single protein.  Read more here

Visualized molecules can make chemistry comprehensible

From the Departmental news letter written by Jes Andersen:

When students experience chemistry as “tough stuff”, it’s often due to the difficulty of fully appreciating molecules that are too small and reactions that are too fast. Chemistry professor Jan Jensen has a method of making chemistry much more visible. Using computer-based simulations and visualisations, he offers his students a more intuitive and less abstract gateway into the field of chemistry. Now, Jensen and his colleagues at the Centre for Computational Molecular Science have received 200,000 kroner from the University’s “Education at its Best” project to develop and diffuse visualisation tools throughout the courses of his fellow chemistry instructors. He is convinced that these tools will make it a bit easier to become a chemist.

“In the oral examinations, I’ve noticed that even those who are quite poor at equations have improved in their ability to explain what is going on at the molecular level,” explains Jensen. In the coming months Jensen and his collaborators in the Centre for Computational Molecular Science will adapt the visualisation tools to courses at the University of Copenhagen.

The hope is that ambassadors from the Centre will ensure that the visualisation methods spread like ripples across water. While Jensen expects to employ student interns to create finished “instruction packages”, he encourages curious professors to sign up for the development phase. “I am confident that this will make it easier to teach chemistry. But here at the beginning, it would be more fun to test ideas in collaboration with professors who actually have a desire to use these tools,” affirms Jensen who is confident that these methods will benefit instructors as well. “I have a few lectures where I do not use the visualisations and I can clearly see that I lose the students’ attention,” concludes Jensen.

Read Jan Jensen’s blog post about visualisation.
Read about the grant.

The Japanese taught him focus… and how to enjoy the weekend

From the Departmental news letter written by Jes Andersen:

Most equate study abroad with academic challenges. But For Kasper Steinmann, the chance to study in the Tokyo suburb of Tsukuba offered lessons which proved to be very human.
Read entire story here.

QM/MM course: notes from week 1

Many questions and discussions during the first lecture.  The course is off to a great start. 

We started by drawing dates and papers and a list of who does what can now be found on the course web page.  You are free to switch dates and papers if you can find someone to switch with.

I also briefly introduced the requirements for the extra credit assignment.  There are now more details here.

Points of discussion not covered by the slides

Size of QM region and QM/MM interactions
The most practical way to gauge the whether the division into QM and MM regions is correct is to enlarge the QM region and see if the results change significantly.  This is rarely done.  If I find some references I will put them here.

Covalent QM/MM boundary
Enlarging the QM/MM region will introduce more cuts across covalent bonds, which could increase the error.  However, the cuts will be further away from the active site, so the effect on relative energies should be smaller, and they should converge as the QM region is increased.

I am not aware of any QM/MM study that does not place the cut between the alpha and beta carbon on the side chain (I would very much like to hear about exceptions).  This is due to the fact that QM/MM boundaries are heavily parameterized, and need to be redone for other types of cuts.  This places restrictions on how much the QM region can be enlarged. 

Slide 12: In the link atom method, there are interactions between the MM charges and the link H atom.  It is not really feasible to remove the “H-part” of the density.

QM/MM packages
Slide 16: Most QM/MM methods consist of a QM and MM program hooked together, sometimes by a third piece of software.  The only “fully integrated” package I know of is Qsite, a commercial program from Schrodinger.  Kasper has put up some more info here.

Tuesday, February 1, 2011

Predicting protein pKa values: problem solved?

Quantum chemistry has reached a point where some gas phase properties of small molecules can be computed more accurately than they can be measured.  For these systems science can say "problem solved", give itself a light pad on the back, and move on to other things - like how to do the same for molecules twice the size.  Or in solution.

I don't believe there is an equivalent success story in biomolecular modeling, but, as I'll argue here, the prediction of the pKa values of ionizable residues in proteins be a close contender!

Jens Erik Nielsen and co-worker recently published a very interesting paper entitled "Remeasuring HEWL pKa values by NMR spectroscopy: Methods, analysis, accuracy, and implications for theoretical pKa calculations" in Proteins.

One of the most interesting implications of the paper, in my opinion, can be found at the bottom of Table III.  Here the authors compare Asp, Glu, and His pKa values in HEWL measured using the NMR chemical shifts of backbone nuclei to those obtained from side chain 13C atoms. 

The RMSD values are 1.3, 1.2. and 0.6 pH units for 15N, 1HN, and 1Hα, respectively (note that the latter could only be obtained for 7 of the 11 residues).  Now compare this to the corresponding values for pKa values predicted with the pKD and PROPKA servers: 0.9 and 0.8. 

Thus, if you are not willing to go through the considerable trouble that is assigning side chain chemical shifts for a protein, you are better off predicting the pKa values than measuring them!  At least for this protein ... and only if you can't live without knowing all the pKa values ... but still.

Nielsen and co-workers estimate the accuracy of the 13C-derived to be accurate to within 0.1 - 0.2 pH units, so, no, the general problem is not solved.  But if more people do as careful a job measuring pKa values as Jens, we may find that we are closer than we think.

Three chemistry bloggers on why you should blog

Noel O'Boyle: Don't keep your cool - share it

Henry Rzepa: Is there a difference between a scientific blog and scientific journal?

Steven Bachrach: Why blog?

Jmol integrated with chemCanvas

From Jmol Users list via Blue Obelisk discussion list (cool idea, but still many bugs):

I though that some people may find interesting the integration of Jmol
with chemCanvas, an open source chemical diagram editor.

chemCanvas is a javascript library (a jQuery plugin) using HTML 5 canvas.

For more information you may visit: