ScotCHEM Computational Chemistry Symposium 2013 – or why I need to learn more Chemistry
Posted: 21 Jun 2013 | 17:05
Last week I attended ScotChem 2013 at the School of Chemistry, University of St Andrews. This two-day event was founded by Carole Morrison (Chemistry, University of Edinburgh) and Tanja van Mourik (Chemistry, University of St Andrews) to bring together computational chemists in Scotland, and I went along to find out how the HPC community is involved in computational chemistry. This was also the first year that the ScotCHEM meeting was held over two days - the first day was focused on a CCP5 workshop on modelling the chemistry and biochemistry of condensed phases. This workshop also aimed to address the underrepresentation of women in Chemistry by showcasing an all-female cast! It certainly is impressive to see so many female computational chemists, as normally I feel surrounded by men!
Unfortunately the first talk, by Prof. Marie Pierre Gaigeot, was cancelled, but Tanja van Mourik stepped in to provide a fascinating talk on 'The mutagenic action of 5-bromouracil (5BrU)' - basically using density function theory (DFT) calculations to help explain how the incorporation of water around 5BrU enables mutations in DNA during replication. 5BrU exists in three tautomeric forms (isomers of organic compounds that can readily interconvert by a chemical reaction), and the three forms frequently interchange, with each tuatomer forming a base pairing with a different neclease (adenine or guanine). When the interchange between tautomers occurs during during DNA replication, mutations can occur. The aim of Dr Mourik's work, which is only just starting out, is to explain how it is possible for this to happen with the aid of computational density functional theory calculation (DFT). The calculation shows that water molecules 'hop' between protons on the BrU, enabling the formation of a different tautomer, reversing the tautomeric preference from what is apparently renegertically favourable. The presence of water in this is crucial as without the water 'hop' the conversion energy is so high it would never occur.
Carole Morrison ("I love working in a department where experimentalists need to be nice to me so I can fill in the gaps in their work") provided the second talk of the day, whereupon I felt thoroughly out of my depth! She spoke about using time-dependent DFT (OK got that bit) to understand a molecular switch which occurs during photo excitation; using UV to close a molecular ring, and visible light to open it (basically there, well mostly). Dr Morrison is using a substitution technique where ground state singlepoint energy calculations and excited state singlepoint energy calculations are combined with a calculation of the open-ring structure, 'substituting' the results, to find peaks in experimental data. The great thing about Dr Morrison's work is she can simulate events and identify structures that her experimental colleagues cannot begin to look for because their molecular crystals are so delicate that they decay after 400 or so on-and-off switches. Dr Morrison can switch the molecule on and off until she runs out of AUs! Her calculations can account for individual diffraction peaks' time-resolved diffraction and help the experimentalists understand their findings. Unfortunately for me, I understood the DFT, but not the chemistry! The fascinating thing about this talk was using a combination of DFT calculations, and subtracting them, to find the diffraction peaks.
These first two talks took me back to the days of my PhD; using DFT to understand molecular structures. However, on day two, I definitely moved out of computational comfort zone and well and truly into chemistry beyond my knowledge base! The plenary speaker on Friday morning was Prof Tim Clark from the University of Portsmouth. Tim educated the audience on the flaws of using simple point charge representations of atoms, ignoring torque and therefore only needing to solve Newton's equations (completely ignoring angular torque and Euler's equations). This of course makes my job as a computational physicist far easier, because I do not have to code up as many equations! However, Prof Clark highlighted that this has led to major flaws in our understanding of non-covalent interactions, which we commonly believe are not directionally biased. I also learned of σ-holes, where charge densities provide viable bonding sites between atoms that would not normally bond. This was definitely a lecture out of my physics comfort zone!
Other talks over the two days focused on particular molecules, simulation of complex molecules at interfaces and categorising enzymes into 'superfamilies'. I learned a lot of chemistry that I did not know before and saw a huge breadth of use for high performance computing, from DFT calculations that I work with every day to pattern matching of three dimensional protein structures.