Highlights from Faraday Discussion: Meeting Challenges in the Analysis of Complex Natural Mixtures, Edinburgh, UK

Against a backdrop of the golden gorse covered volcano (extinct) illuminated with strong sunshine (Fig. 1), 98 delegates from around the world gathered in the John McIntyre Conference Centre, The University of Edinburgh, to discuss the most challenging of analytical problems: complex mixtures. Complex natural mixtures range from petrochemicals, food, drink and bodily fluids to natural organic matter found in all ecosystems. The common feature to all these mixtures is that they consist of hundreds to 10 000s of molecules of varying chemistries that challenge all analytical techniques. At the same time, as the sample complexity increases, more powerful tools to mine the data to form robust, meaningful conclusions are required. To this end, chemometric and other computational methods are being utilised more and more. Given the scale of the challenge, a multi-disciplinary effort is required to tackle complex natural mixtures and, for the first time, a deliberate fusion of expertise in these diverse fields was an overarching principle behind the program selection for this Faraday Discussion.

Fig. 1 Gorse covered Arthur’s Seat as seen from the John McIntyre Conference Centre. (Photo by Justin van der Hooft.)

Upon arrival, delegates picked up their badges and conference booklet from the RSC team. To celebrate the International Year of the Periodic Table, everyone was given a lanyard decorated with a rather colourful periodic table, a token met with a smile by many. Beside the registration table, one's eyes were drawn to the large timeline of the Faraday Discussions underlining the privilege and importance of contributing to these meetings.

The first Faraday Discussion meeting was held on the 29th January 1907 where delegates discussed osmotic pressure. Since then the meetings went from one to around eight per year, with the scope expanding from topics of pure chemical physics to those at the interface of chemistry, physics and biology. The Faraday Discussions have a unique format - papers are submitted by speakers prior to the meeting and read by those planning to attend. The pre-conference preparation allows time at the conference for a brief introduction to each paper followed by a comprehensive discussion. In this way, not only can the work presented be challenged, but at the same time a multi-disciplinary view can inspire new directions.

After a buffet lunch in the Pentland West room, delegates gathered in the lecture hall for the conference opening. The Conference Chair, Dušan Uhrín, delighted the audience with an introduction to the City of Edinburgh and highlighted what to see if time allows (Fig. 2). May is one of the nicest times to see Edinburgh, with the gardens and parks in full bloom. He introduced the conference committee (Mark Barrow, Timothy Ebbels, Ruth Godfrey, Donald Jones and Mathias Nilsson), the RSC team (Susan Wetherby and Sarah Thirkell) and helpers (Richard York and Marc Stockwell) before recapitulating the format of the meeting.

Fig. 2 Prof. Dušan Uhrín, Conference Chair, opening the meeting with an introduction to the City of Edinburgh. (Photo by Gianluca Trifirò.)

Each speaker is allowed to present for 5 minutes followed by a 15 minute discussion, initially directed at the individual paper, followed by a more general discussion to all session presenters. Each question is recorded diligently by the RSC editing team and the question and response published in the conference proceedings.

The RSC publishing editors (Suzanne Howson and Ella Wren) followed with an explanation of how to respond to the recorded questions and provide answers for the publishing process. In addition, they highlighted the Twitter hashtags for the conference #fdmixtures and #fd2019.

Opening lecture

The Conference Chair introduced the plenary speaker as ‘one front runner in applying multiple analytical tools to tackle complex mixtures’. Philippe Schmitt-Kopplin, Helmholtz Zentrum, München, Germany, mesmerised the audience with his opening lecture entitled “Systems Chemical Analytics” (Fig. 3). He set the framework of the conference by defining complex natural mixtures. He challenged the idea of complexity starting with an image of the Rosetta stone. The message was clear: what looks complex may in fact be trying to simplify something. He solicited thoughts of others on what is complexity, providing some memorable quotes: ‘Complexity is what we see when we open the eyes – complexity is beauty’ (Regis Gaugeon).

Fig. 3 Prof. Philippe Schmitt-Kopplin giving the opening lecture. (Photo by Will Kew.)

The need for an integrated approach which combines separation science, spectroscopic, spectrometric and mathematic (chemometric) techniques to clarify samples according to chemical composition relationships was called upon. At the same time he acknowledged that analysis time vs. resolution vs. cost determine our efforts in one area or another, depending on what samples we are looking at or what question we are trying to answer. Setting the conference spirit, a number of examples of how big questions regarding the complex natural mixtures, the metabolome and holobiome, can be addressed by such an integrated approach were presented - from the chemistry of marine holobionts, Kendrick-Analogous Network Visualisation of Fourier transform-ion cyclotron resonance-mass spectrometry (FT-ICR-MS) data of secondary aerosols, the non-targeted analysis of wine aged in wood from different regions of France, the diversity of compounds produced from the thermal treatment of simple starting materials to understand the inner workings of the Maillard reaction, to dissolved organic matter (DOM) and the introduction of chemical dark matter. The lecture finished setting some major challenges for the field of complex mixtures; the need to ensure good quality data, tools to integrate data collected 5–10 years ago with today's data, automated procedures for thousands of samples and approaches to handle big data to answer the questions appropriately.

Session 1. Dealing with complexity (Chaired by Mark Barrow and Ruth Godfrey)

Ryan Rogers (Future Fuels Institute, USA) kick started the first session with a mind-blowing presentation of his work on ‘Combating selective ionization in high resolution mass spectra characterisation of complex mixtures’. A point made clear by the 21T data of crude oil mixtures; while the direct ‘dilute and shoot’ method has become the standard for FT-ICR-MS analysis of complex mixtures, selective ionisation may mean that we are ‘seeing’ only 10–20% of the whole sample. The approach to solve this problem was illustrated by chromatographic fractionation prior to ionisation. Using a mixture of fractionation methods allowed a substantial increase in the number of ions produced, particularly those of higher molecular weight. It is clear that the determination of all compounds in very complex mixtures is beyond the capability of current FT-ICR-MS methods however, it was acknowledged that the complete characterisation is not always needed to answer the specific research question asked.

Continuing on the theme of what we are not ‘seeing’ with high resolution mass spectrometry (HRMS), Jeffery Hawkes (Uppsala University, Sweden) presented the paper ‘Complexity of dissolved organic matter in the molecular size dimension: insights from coupled size exclusion chromatography-mass spectrometry’. Dissolved organic matter (DOM) is a form of organic matter in natural water, whether that is rivers, oceans, lakes or the pore waters of soil. The mass distribution generally reported by FT-ICR-MS is between 200–800 Da, while other methods such as size exclusion indicate sizes much larger than 800 Da. This difference has been traditionally regarded as evidence of either aggregates, or due to selective ionisation. Jeffrey, however, illustrated using size exclusion chromatography coupled with an online UV-diode array detector and electrospray ionisation-HRMS the existence of large (1000 to >2000 Da) UV-active DOM molecules that were simply not ionised. From the analysis this pool of compounds is not composed of aggregates and is phenolic in nature. This study demonstrates that care should be taken in combining the results from optical and mass spectrometry methods for characterising DOM as they are detecting different types and sizes of molecules present.

After a 30 minute coffee/tea/biscuit break, Peter Schoenmakers, University of Amsterdam, The Netherlands, gave his ‘Perspectives on the future of multidimensional platforms’, which offers a comprehensive insight into the current state of 2D liquid chromatography (LC). 2D LC can either be performed by subjecting the entire eluent of the first column to a second column (LC × LC) or in ‘heart-cut’ mode (LC–LC) where one or selected fractions are directed to the second column for separation. While 2D LC increases the separation power, several challenges have prevented its routine use. Namely, reduced detection sensitivity, solvent incompatibility and complex method development. Solutions to these challenges were suggested such as the use of active modulation techniques between the LC columns. For example, stationary phase-assisted modulation involves trapping the analytes after elution from the first column, which are then washed for purification and desorbed onto a second column for further separation. A number of applications were summarised, from the use of reactive modulators such as enzymes for drug-delivery systems or light for dye degradation products, to the use of spatial 2D LC where the stationary phase could be programed in a ‘pantoflutographic’ manner to allow concurrent 2D separation. The addition of modulators, however, makes the method development more complex and the creation of automated method development routines are therefore essential. The authors presented their first iteration of a program for automatic peak tracking required for such automated method development and discussed the role machine learning could have for advances in this area.

Keeping on the subject of multi-dimensional platforms, Danilo Sciarrone, University of Messina, Italy, presented the ‘Collection and identification of an unknown component from Eugenia essential oil exploiting a multidimensional preparative three-GC system employing apolar, mid-polar and ionic stationary phases.’ Using wide-bore columns (which allow micro-milligram sample injection) and three stationary phases, with only five runs in heart-cut mode, a 95% pure fraction of an unknown compound that contributed to 35% of the essential oil, was obtained via this gas chromatography (GC) approach. Interestingly, the unknown compound was simply collected in a tube by condensation for immediate dissolution in a solvent for ‘standard’ NMR analysis leading to the identification of 6-ethenyl-6-methyl-3,5-di(prop-1-en-2-yl)cyclohex-2-en-1-one.

The first session came to a close with the last presentation by Dumitru Duca, University of Lille, France, who gave his insights ‘On the benefits of using multivariate analysis in mass spectrometric studies of combustion-generated aerosols.’ The surface chemistry of soot generated from three different sources was investigated by two MS methods; secondary ion mass spectrometry (SIMS) and two-step laser desorption ionisation (LDI) mass spectrometry. In order to determine if different engine regimes had an effect on gasoline derived soot, principle component analysis (PCA) was used showing a clear separation between soot produced from a normal working engine and a malfunctioning engine. Furthermore, a clear separation between non-optimised regimes caused by hydrogen rich hydrocarbons was apparent. Hierarchical cluster analysis of all samples allowed clear visualisation of sample clusters underpinned by differences in chemical composition. The hope here is to pave the way for standardised approaches to soot measurement and data analysis.

Poster session

After completion of session 1, lightning presentations were held for all 26 posters. Each presenter had one slide and 45 seconds to entice the delegates to come see their poster. A fantastic achievement by all, with some only needing 30 seconds to get the message across! As soon as the last presentation was given, everyone proceeded into the Pentland West room for drinks and the poster session. The posters covered a wide range of applications; ion mobility MS (IM-MS), novel NMR experiments and processing and analysis methods, multiple ionisation modes coupled to MS including LDI, MS processing methods, hyphenated methods such as supercritical fluid chromatography coupled to MS or GC-IM-MS or LC-IMS-Q-TOF, combinations of multiple techniques and a few examples of statistical methods. These tools were applied to a diverse range of complex mixtures from soils, proteins, crude oils, fracking oils, metabolites carbohydrates, fermentation products, aerosols, semi-volatile compounds, beverages, algae-derived bio-crudes, insecticides, disinfection by-products, dissolved organic matter and Maillard reaction products. The poster session was a great opportunity for the postgraduate students to discuss their work with experts from different fields. The networking continued after the poster session closed as delegates left to explore the City of Edinburgh's many restaurants for dinner.

Session 2. High resolution techniques (Chaired by Don Jones and Mathias Nilsson)

The second day promised to be another sunshine filled day. Delegates arrived early to be greeted by tea/coffee and mini pastries.

Ljiljana Paša-Tolić, Pacific Northwest National Laboratory, USA, started the session with a presentation entitled ‘Online supercritical fluid extraction mass spectrometry (SFE-LC-FTMS) for sensitive characterisation of soil organic matter.’ The use of online SFE-micro-Solid Phase Extraction (SPE)-nanoLC coupled to an Orbitrap and FT-ICR-MS allowed nearly twice the number of peaks to be obtained from intact soils compared to standard direct infusion methods. Part of the reason for this substantial increase was the use of LC to reduce ion suppression, and the ability of supercritical-CO₂ to penetrate soil pores, improving the extraction. Of those detected, a substantial number could not be assigned (around 50%) showing that a large proportion of these molecules accessible by SFE are not in assignment databases nor contain elements within the assignment algorithms. The ability to obtain a large coverage of compounds from a minimal amount of soil paves the way to use this method for spatial characterisation of organic matter from soils.

Alexander Zherebker, Skolkovo Institute of Science and Technology, Russia, presented a paper entitled ‘Structural investigation of coal humic substances by selective isotopic exchange and high-resolution mass spectrometry.’ Here the combination of selective liquid-phase H/D exchange and FT-ICR-MS allowed the counting of exchangeable protons in molecules contained in humic substance samples extracted from three different coals. The coal samples shared over 2000 similar molecular formulas, with 500–700 of aromatic in nature. Many of these aromatic molecules had the same molecular formula but differed in terms of side chains. A correlation between the number of exchangeable protons and double bond equivalents for these molecules prompted chemical transformations to be suggested that give insight into the lignin humification process.

Nicholle Bell, University of Edinburgh, UK, presented ‘Reduced dimensionality hyphenated NMR experiments for the structural determination of compounds in mixtures’. To determine the structure of molecules in complex mixtures by NMR, Heteronuclear Single Quantum Correlation (HSQC)-hyphenated experiments that separate proton correlations that have a particular C–H pair source are typically required. If two molecules have the same ¹³C chemical shift for a given C–H pair, multiple correlations will appear on the same line, with no way of identifying which molecules they belong to. A solution to this was presented in the form of 2D NMR through-bond and through-space experiments that use the unique combination of the C–H carbon and proton chemical shifts to separate out the carbon chemical shifts and hence allow the different molecules to be distinguished. The power of these so-called reduced dimensionality experiments was demonstrated on the structural characterisation of green tea carbohydrates.

After coffee/tea/more pastries, Dan Stræk, University of Copenhagen, Denmark, presented his contribution entitled ‘Unravelling the complexity of complex mixtures by combining high resolution pharmacological, analytical and spectroscopic techniques: antidiabetic constituents in Chinese medicinal plants.’ The paper demonstrated methodology for the systematic pharmacological screening and subsequent structural identification of inhibitory molecules from Chinese medicinal plant extracts. Single concentration inhibition percentages of three important targets for the management of type 2 diabetes were used in combination with HPLC-micrOTOF-Q MS and off-line NMR. This protocol allowed fast identification of natural products that had one or more bioactivities against Type 2 diabetes.

Further developments of chromatographic technology were introduced by Fleur van Zelst with a presentation titled ‘Characterising polar compounds using supercritical fluid chromatography-nuclear magnetic resonance spectroscopy (SFC-NMR).’ Specifically this detailed the hyphenation of SFC and NMR along with novel sample injection modes using a sCO₂ flush solvent. This was found to increase the concentration of the sample plug during injection due to sCO₂ expansion and was applied for non-polar compounds within patient samples to increase the sensitivity of NMR to better meet trace analysis. However, following the true purpose of the Faraday Discussions, the reasoning why SFC has struggled to cement its place for routine analysis was debated and how some of these challenges have or are being addressed.

The last contribution to this discussion was Pedro Lameiras, CNRS, France, who presented ‘Polar mixture analysis by NMR under spin diffusion conditions in viscous sucrose solution and agarose gel’. The use of viscous solutions or gels causes the tumbling rate of molecules to slow down which, in turn, allows magnetisation between protons to be shared in the same molecule through intramolecular spin diffusion. In this work structure determination of four dipeptides using 2D NOESY and 2D HSQC-NOESY was compared using mixed solvents/gels, namely water/D₂O:sucrose and D₂O:agarose. These two media were chosen due to their common availability and low cost. The use of sucrose however led to spectral interference from the signals from sucrose itself but a mitigating measure using selective excitation was also presented. Agarose on the other hand did not present signals in the NMR spectra of dipeptides providing a more ‘simple’ analytical approach, however it has a smaller operating temperature range, limiting the application of this medium. On the whole, the sucrose based solvent proved the most suitable for the analysis of dipeptides, with the authors noting that agarose gel appears suitable for rigid dipeptides only.

Session 3. Data mining and visualisation (Chaired by Timothy Ebbels and Mark Barrow)

After a buffet lunch, everyone gathered once more to start the discussion on how to handle the data we get from various analytical techniques.

First to present was Johan Trygg, Umeå University, Sweden, discussing his paper ‘Joint multiblock analysis of biological data-multiomics malaria study’. This involved a multiblock approach, unsupervised integration of multiple data sets, for statistics and how relationships between data can be visualised. The method was used to integrate the analysis of lipidomic, metabolomics and oxylipin data obtained from plasma samples from children infected by malaria. The potential challenges to this approach were also described; these included the effect of data transformation on the correlations derived from the statistical process and the use of the multiblock approach to compensate for ‘missing’ data. Finally he proposed that techniques such as PCA and multiblock could be used to screen the sample followed by a targeted statistical approach to provide enhanced pattern recognition as a diagnostic tool.

Following on from pattern recognition at a macro level, Justin van der Hooft of Wageninen University, The Netherlands, introduced a novel molecular elucidation approach using mass motifs titled ‘Deciphering complex metabolite mixtures by unsupervised and supervised substructure discovery and semi-automated annotation from MS/MS spectra’. This paper essentially discussed an alternative way of characterising molecules using substructures produced by MS fragmentation and storing them in a database for use in compound identification. The discussion also addressed limitations of the algorithm and its use with different instrumentation, e.g. high and low resolution data (by changing the bin size for the dataset) and the probability values for the model in motif searching with data acquired on different instruments.

Tim Rudd, National Institute for Biological Standards and Control, UK, gave the last contribution before the tea break regarding ‘Multivariate analysis applied to complex biological medicines’. The subject of his presentation was heparin, used world-wide as an anticoagulant drug. Heparin has traditionally been extracted from intestinal muscosa of mainly porcine animals however, in 2008 there was a global health emergency when heparin was adulterated, leading to a demand for routine monitoring using high resolution techniques such as NMR. When dealing with a large number of samples, and hence spectra, multivariate tools such as PCA are crucial, and this methodology has been used to successfully identify contaminated heparin samples. However the market for new additional supplies of heparin brings the need for protocols to help differentiate animal sources. This study proposed to address this requirement by performing PCA on 2D NMR spectra of heparin which allowed the differentiation of heparin bovine, ovine and porcine sources.

We continued the session after afternoon tea, coffee and muffins with Jeroen Jansen, Radboud University, The Netherlands, who presented ‘Resolving complex hierarchies in chemical mixtures: How chemometrics may serve in understanding the immune system.’ This paper evaluated four chemometric approaches to chemically characterise cells using flow cytometry data. Of these, PCA showed potential as an initial approach to clustering different types of cells based on their surface chemistry, revealing different cell populations that were not previously visible. This work also highlighted inherent dangers of using existing chemometric techniques such as sparsity for this data type in limiting the hierarchical detail that can be obtained.

Ricardo Moreira Borges, Federal University of Rio de Janeiro, Brazil, presented the last paper of the session ‘An integrated approach for mixture analysis using MS and NMR techniques’. The idea behind this paper was to improve the identification of known and unknown molecules in natural product mixtures. Unlike metabolomics, where compounds are well-characterised and extensive databases exist, the diversity of natural products means that compounds extracted are not found collectively in one database. Therefore methods are required to quickly identify compounds and prevent replicated results. A free to use software pipeline was therefore created that uses 2D NMR to confirm MS identification and demonstrated on a model mixture and an alkaloid rich plant extract. The method starts with LC-ESI-MS/MS molecular network dereplication using the Global Natural Products Social Molecular Networking tool with open-access data. The next stage involves searching Pubmed for compounds of similar structure to the lead compounds. The Heteronuclear Multiple Bond Correlation (HMBC) and HSQC spectra of these compounds are then predicted and compared with the mixture spectra. The similarity is calculated with the output being a desirable high confidence identification.

Conference dinner

At 6 pm, delegates walked across campus to the South Hall where a drinks reception was held in the evening sunshine before everyone took their seats in the grand hall for a three course meal. The South Hall traditionally hosts Burns supper and, as rightly said by the Conference Chair, Dušan Uhrín: ‘we were all about to take a cup of kindness later’.

Scottish hot smoked salmon was followed by roast duck breast with caramelised apple and spring greens, sweet potato fondant, glazed carrots, shallot and calvados jus. After dinner a very lemony tart with lemon and orange cream and hazelnut sable biscuit was accompanied with coffee/tea and Scottish tablet.

After dinner, Will Kew, Pacific Northwestern National Laboratory, USA wetted everyone's appetite for an after dinner dram with a fact filled talk entitled ‘Exploring the (Complex) Chemistry of Scotch Whisky’ (Fig. 4). Following Will, and on behalf of the president of the Faraday Discussions, Prof. Eleanor Campbell presented the poster prizes jointly to Gianluca Trifirò, University of Edinburgh, and Sheri Murrell, University of South Wales. She then commanded the Faraday Discussion traditional ceremony of the Loving Cup. A 18th century silver cup made by a female silversmith which is toasted ‘in piam memoriam of G. S. Marlow and Angela and Tony Fish’. For the first time in the Faraday Discussion meetings’ history, the loving cup ceremony was not performed with port but with a Single Malt Scotch whisky instead. Ten quaichs, Scottish two handed drinking vessels, inscribed suitably for this meeting with the words: ‘Complex Mixtures Only’ were used (Fig. 4).

Fig. 4 Top: Presentation of a rather tasty complex mixture, Scotch whisky. (Photo by Laurent Duval #BIFROST.) Bottom: “Complex Mixtures Only” Quaich used in place of the Faraday Discussions Loving Cup. (Photo by Dan McGill.)

Each table had a quaich filled with the aqua vitae which was sipped and passed around the table following the strict bowing tradition. An appropriate culmination of a Faraday Discussions conference dinner held in sunny Edinburgh. See #fdmixtures for a brief video of the ceremony.

Session 4. Future challenges and new approaches (Chaired by Mathias Nilsson and Dušan Uhrín)

After a fantastic conference dinner and after party, everyone trickled in to the John McIntyre Centre, grabbed a pastry and coffee or water to start the morning discussion session.

Ronald Soong, University of Toronto, Canada, presented first with the paper entitled ‘Focusing on “the important” through targeted NMR experiments: An example of selective ¹³C–¹²C bond detection in complex mixtures’. This work offers an approach to monitor the fate and reactivity of carbon in environmental and biological samples using a selective ¹³C–¹²C bond forming NMR experiment. The experiment filters out the intact ¹²C and ¹³C signals leaving only signals from structures which have a new ¹²C–¹³C bond. A number of examples were given, for example the ¹³C-glucose fermentation in baker's yeast allowed the confirmation the precise mechanism for the formation of ethanol. Ronald finished his presentation with a fascinating example of the experiment being used for an in vivo study of Daphnia magna fed with ¹³C-algae then ¹²C-algae using a NMR flow system. The experiment demonstrated how Daphnia modify lipids from the algae using carbon in their own system, a process not well-understood.

Dan McGill, Imperial College London, UK, continued the NMR based studies with the application of novel SPE-NMR protocols for metabolic profiling of human urine. It was interesting to hear that the urine metabolite database has over 4000 metabolites but ∼2000 are not detected in human urine to date, a fact that enthused delegates to look for new methods to identify metabolites in biological fluids. This paper tested the ability of 21 SPE protocols (varying cartridge type and condition) to separate components of human urine for NMR analysis. Otherwise obscured eluents were enriched by using certain cartridge types, pH and conditions.

Francisco Fernandez-Lima, Florida International University, USA, finished this session with ‘Understanding structural complexity of dissolved organic matter: Isomeric diversity’. The combined use of trapped ion mobility (TIMS) and FT-ICR-MS as well as FT-ICR-MS/MS enabled the characterisation of isomers making up each of the individual nominal masses observed in HRMS spectra of DOM. Each peak in the high-resolution mass spectrum has been known to exist as a number of isomers although their observation has not been possible using FT-ICR-MS alone. From one nominal mass over 250 structural isomers were reported highlighting the incredible complexity of these mixtures. The paper demonstrated the need for TIMS for a full characterisation of such complex mixtures.

After a few last biscuits, everyone gathered for the last session of the conference, which by pure coincidence was a French affair. To start, Marc-André Delsuc, University of Strasbourg, France, presented his work titled ‘Automatised pharamacophoric deconvolution of plant extracts-application to Cinchona bark crude extract’. This concerned dereplication strategies for natural product elucidation and used a statistical approach for simulating NMR spectra for dereplication. Pleasingly for fellow scientists involved in structural elucidation and biomarker discovery, this statistical tool is available as an open access ‘plasmodesma’ web-based ‘app’, offering the opportunity to be used for an ever increasing range of project work.

Caroline Chaux and Laetitia Shintu, Aix-Marseille University, France, presented work titled ‘Challenges in the decomposition of 2D NMR spectra of mixtures of small molecules’. This involved a discussion of a new mathematical algorithm for NMR data to detangle the multiple components in complex mixtures and provide better estimates of analyte concentration. The algorithm was shown to work well for 1D and 2D spectra of simulated mixtures. For real mixtures the algorithm did not reach the same performance evaluation indices, however the concentrations were better estimated from 2D vs. 1D spectra. The work was highly multi-disciplinary requiring both mathematics and analytical science expertise; it was acknowledged that this was not without its difficulties during the early stages of the project in which each party had to understand each other's scientific language.

The final research presentation of the conference was delivered by Caroline Gauchotte-Lindsay of the University of Glasgow, UK. This was titled ‘Systems biology approach to elucidation of contaminants biodegradation in complex samples-integration of high-resolution and molecular tools’ which examined the link between the chemical signature of environmental pollutants and the effect on the microbial community. This preliminary study shows that the microbial community is dependent on the presence of chemical contaminants, emphasizing the need to couple these diverse datasets for the rational design of remediation strategies. The analytical tools were discussed, including the role of GC × GC for compound identification and the extension of this capability for more polar molecules using derivatization reagents. The challenges for using this approach were also explained and highlighted that the peak detection process was the rate limiting step in the workflow, requiring further development for high(er)-throughput work.

Closing lecture and final remarks

A very entertaining closing lecture was presented by Roy Goodacre, University of Liverpool, UK (Fig. 5). Illustrated using the Blind men and the elephant poem, Roy warned that our preconceived notions of complex mixtures could lead to misconceptions.

Fig. 5 Prof. Roy Goodacre giving the closing lecture. (Photo by Will Kew.)

He reminded us that even the slightest molecular change can give rise to a complexity and that something which seems not so complex could in fact contain more complexity than we anticipate, a point made clear by the example of Chicken Tikka Masala (Roy's favourite dish). This may be thought of as just chicken with sauce however there over 15 ingredients in varying amounts that contribute to the taste of the dish. Roy's lecture presented a lucid summary of the conference including whisky, spinning shrimp, dark matter and a reminder that we need to talk, listen and collaborate to move forward.

After the lecture, the Conference Chair thanked everyone for being active in the discussions, the scientific committee, RSC staff and conference helpers. As a kind touch each of the invited speakers got to take home a Quaich to remember the event and to share few “Complex mixtures only” with friends and family. Delegates left the Pentland West for the last time, grabbed a lunch box and dispersed out to explore the city in the warm sunshine or headed for the airport or train station for the journey back home.

---