Monthly Archives: March 2013

Webinar with Dr. Erik Lindhal – NVIDIA+GROMACS


Thanks to Devang Sachdev from NVIDIA for bringing this webinar to my attention.

The future of computational chemistry seems to be written in CUDA for GPU’s specially when it comes to Molecular Dynamics; as such, NVIDIA has gone through great lengths into introducing scientific computing methods for GPU’s. I still have a pending review of a test drive that people at NVIDIA and EXXACTCORP kindly allowed me to run but that is the topic of the next post.

Next Thursday, April 4th, 2013 from 9:00 AM – 10:00 AM Pacific Standard Time there will be a webinar in which Dr. Erik Lindhal at Stockholm University and NVIDIA will discuss latest GPU-acceleration technologies available to GROMACS users; more specifically the latest accelerated version of GROMACS 4.6, which features are supported, it’s installation and use, and how it performs with latest NVIDIA Kepler GPUs.

Register here: http://goo.gl/0HtqJ

Please register and check your local timezone to avoid delays. I will register as soon as I finish typing this. Thanks once again to Devang Sachdev for all his help, patience and trust in this forum.

Advertisements

Article in ‘Ciencia y Desarrollo’ (Science and Development)


Here is a link to an article I was invited to write by my good old friend, Dr. Eddie López-Honorato from CINVESTAV – Saltillo; Mexico, for the latest issue of the journal ‘Ciencia y Desarrollo’ (Science and Development) to which he was a guest editor. ‘Ciencia y Desarrollo’ is a popular science magazine edited by the National Council for Science and Technology (CONACyT) of which I’ve blogged before.

This magazine is intended for people interested in science with a general knowledge of it but not necessarily specialized in any field. With that in mind, I decided to write about the power of computational chemistry in predicting some phenomena while shedding light in certain aspects of chemistry that are not that readily available through experiments. The article is titled ‘Chemistry without flasks: Simulating chemical reactions‘. The link will take you to the magazine’s website which is in Spanish, as is the article itself, and only to the first page; so, below I translated the piece for anyone who could be interested in reading it (Hope I’m not infringing any copyright laws!). Don’t forget to also check out Dr. López-Honorato’s blog on nuclear energy research and the development of materials for nuclear waste containment! Encourage him to blog more often by liking and following his blog.

 

Chemistry without flasks?
Typically we think of a chemist as a scientist who, dressed in a white robe and protected with safety glasses and latex gloves, busily working within a laboratory, surrounded by measurement equipment, glassware and bottles with colored substances; pours one substance onto other substance, transforming them into new substances while noting that the chemical reaction occurs through color changes, heat release , perhaps gas, and occasionally even an explosion.
Thus chemistry, the study of the material processing involves active experimentation to accomplish chemical reactions subsequently confirmed, although indirectly, that the changes have been conducted in the microscopic world, moreover, in the molecular and atomic world. The chemist plans these changes based on the knowledge he has of the chemical properties of the substances of which he started and, like any other substance, are due to its molecular structure, i.e., the spatial arrangement of the atoms that form it.
 
Under this archetypal image just posed, then it’s at least funny to think that there is a branch of chemistry named Theoretical Chemistry.

What is theoretical chemistry?
Theoretical chemistry is a kind of bridge between chemistry and physics; using laws and equations that govern the subatomic world, to calculate the molecular structure of a substance, more specifically calculate the distribution of electrons surrounding the molecule forming a cloud, which interact with the electron cloud of another molecule to form a new substance. It is based on the knowledge of the electron density cloud or we can understand and predict the chemical properties of any substance. We can then define theoretical chemistry as the set of physical theories that describe the distribution and properties of the electron cloud belonging to a molecule, in this particular mathematical description we call electronic structure and this is the starting point for descriptions and chemical predictions.

What is it good for?
Through theoretical chemistry we can find answers to fundamental questions about the structure of matter. Consider a molecule of water, which has the chemical formula H2O. This formula implies that there are two hydrogen atoms attached to an oxygen atom But what spacial structure does a water molecule have? The simplest geometry it could take would be a linear structure, in which the angle formed by the three atoms is 180 °. However, the water molecule has an angle of 109 °, far from a linear structure. In Figure 2 we can see the result of the calculation of the electronic structure of H2O, it observed that the electron cloud that exists on the oxygen atom also has a place in space and thus push the hydrogen atoms bringing them together instead of allowing them to take a  more comfortable conformation.
 
Figure 2. Oxygen remaining electrons (red cloud around the oxygen atom) that are not chemically push the hydrogen atoms towards each other.

The industrial area currently impacted by the application of theoretical chemistry is the pharmacist, as they generate a new drug involves significant investment in financial and human resources, so predicting the properties of a molecule with pharmacological activity before synthesizing is highly attractive. Therefore it has been generated within the theoretical chemistry field, otherwise known as branch Rational Drug Design.
Drugs acting on our organism when active molecules interact directly with the various proteins which are distributed in the tissue cells. If the structure of the protein is known and we attack is known also a drug which acts on it, then we can design similar drugs having greater efficacy in the treatment of diseases. But it is not only fit one molecule to another, but to calculate the energy of interaction, the energy of dissolution and the probability that this interaction can be observed experimentally (Figure 3). The calculation of the interaction energy between the drug and the protein tells how strongly attract each other, a weak attraction drug will result in a low efficiency, while a greater attraction involve a more effective drug.
 
How do you calculate a molecule?

All matter exists in the universe is made of atoms, which in turn are composed of a nucleus of protons and neutrons surrounded by a cloud of electrons. When two or more atoms combine to form a molecule combining do their electron clouds and how do these combinations are best described by the equations of quantum mechanics, the branch of physics that describes the behavior of the subatomic world. However, due to its complexity, the equations of quantum mechanics can only be accurately resolved in the simplest cases such as the hydrogen atom, which consists of a single electron orbiting a proton. We must therefore resort to a range of methods and approaches to tackle cases of chemical interest and even biological.
For years the only available computers could solve the approximate equations for small molecules, no later than thirty atoms, which which can be interesting, but not entirely useful. Today modern supercomputing equipment (which may amount to up to tens or even hundreds of powerful computers connected together to work cooperatively) allow us to make models with hundreds of atoms molecules such as proteins or DNA fragments.
While the software available to perform these calculations is developed continuously for the last thirty years has been the progress in the design of computer systems able to perform thousands of operations per second the cornerstone that has made the theoretical chemistry a predictive tool commonly used. Today the branch known as Molecular Dynamics, which studies the interactions between molecules over time, has benefited from the development of the latest game consoles, as their processors, known as graphics processing units (GPUs , for its acronym in English) are able to perform calculations in parallel: Many of the images seen in our video games are actually calculated, not animated, this means that the console must calculate how to answer each item on the screen According to each stimulus we introduce. Conversely, if the images were animated, the answers would be always the same and the game would become unrealistic. Each game event should be calculated almost immediately to maintain its fluidity and emotion, in such a way that these GPUs have to be able to perform several mathematical operations simultaneously.
Traditionally molecular dynamics is based on the equations of classical physics, which only see the time evolution of molecules like solid objects collide, hundreds of molecules floating in water or other solvent. With the advent of GPUs can include dynamic calculating the electronic structure so we can peek into biological processes such as DNA replication or the passage of nutrients through a channel protein embedded in the membrane of a cell.
Conclusions
Since the fundamental understanding of the distribution of electrons in a molecule, its structure and properties to rational drug design, new materials based on molecular modeling theoretical chemistry is a powerful tool which is constantly progressing. The development of computer systems increasingly powerful detail allows us to meet the electronic processes involved in a chemical reaction while we can predict the real-time progress of molecular transformations. All this brings us ever closer to the dream of modern alchemists: transform matter to obtain substances with properties designed to pleasure.
In the nineteenth century, the American philosopher Ralph Waldo Emerson, wrote: “Chemistry was born from the dream of the alchemists to turn cheap metals into gold. By failing to do so, they have accomplished much more important things. ” And yes. Today we delve into the innermost secrets of nature not only to understand how it works but also to modify its operation on our behalf.

Natural Bond Orbitals (NBO) Visualization with Chemcraft


It’s been a long time since I last posted something and so many things have happened in our research group! I should catch up with them in short but times have just been quite hectic.

I’m glad to publicly thank Prof. Frank Weinhold’s gesture to include this blog in the bibliography section of the new NBO6.0 website under the NBO-Related Websites tab.

Here is a contribution from Igor Marques at the University of Aveiro in Portugal (Group Website); the original text can be found as a comment in the original NBO Visualization post but it is pretty much the same thing you can find in this post. Here is a link to Chemcraft’s website. Thanks for sharing this, Igor!

=> Examples provided by Igor Marques used Chemcraft Version 1.7, build 365 <=

In the Gaussian input, with the NBORead option included under the population keyword, we should include the PLOT option as illustrated below. The gfoldprint keyword will print the basis set to the output file in the old G03 format. Some visualization programs require a certain format of the basis set to be printed to the output file in order to plot orbitals and other surfaces like the electron density; therefore, if you want to play safe, use gfoldprint, gfprint and gfinput in the same line. gfprint will print the basis set as a list but in the new G09 format, whereas gfinput will print the basis set using Gaussian’s own input format. (The used level of theory and number of shared processors are shown as illustrations only; also the Opt keyword is not fundamental to the visualization of the NBO’s)

%chk=filename.chk
%nprocshared=8

#P b3lyp/6-311++g** Opt pop=(full,nboread) gfoldprint 
filename

0 1
molecular coordinates
$NBO BNDIDX PLOT $END

this will generate files from *.31 to *.41
For the visualization of NBOs, you’ll need FILE.31 and FILE.37. Open FILE.31 from chemcraft. It will automatically detect FILE.37 (if in the same directory).

Tools > Orbitals > Render molecular orbitals

select the NBOs of interest (whcih are in the same order of the output),

Adjust settings > OK

On the left side of the window, select the NBO of interest and then click on ‘show isosurface’. Adjust the remaining settings. To represent another orbital, click on ‘keep this surface’ and then select another orbital from the rendered set and follow the previous steps.

Some Considerations:

> It’s possible to open a formated checkpoint file, containing the NBOs, in chemcraft.
Gaussian input:

%Chk=filename.chk
%nprocshared=4
#P b3lyp/6-311++g** Opt pop=(full,nboread,savenbo) gfoldprint 
filename

0 1
molecular coordinates
$NBO BNDIDX $END

the procedure is identical, but it is only necessary to read the *fchk file and then render the desired orbitals.
However, two problems might arise:
a) Orbitals in the checkpoint are reordered, thus requiring some careful inspection of the output.
b) Sometimes, for a larger molecule, the checkpoint might not be properly saved and the Gaussian job (as previously reported – http://goo.gl/DrSgA ) will end with:

Failed in SchOr1 in NBStor.
Error termination via Lnk1e in /data/programs/g09/l607.exe at Wed Mar 6 15:27:33 2013.

****

As usual, thanks to all for reading/commenting/rating this and other posts in this blog!

%d bloggers like this: