skip to content

Department of Plant Sciences



Supervisor: Dr Andrew Tanentzap

Co-supervisor: Prof Thorsten Dittmar (University of Oldenburg)


Brief Summary

Invisible to the naked eye lies a tremendous diversity of organic molecules that shapes the world's biogeochemical cycles.


Importance of Research

Dissolved organic matter (DOM) is central to the functioning of freshwaters that support Earth's life systems, such as by changing the attenuation of sunlight, altering contaminant toxicity, and increasing nutrient pools.
Freshwater DOM is also one of the largest natural fluxes in the global carbon cycle. However, predicting the composition of DOM has been challenging given its tremendous molecular complexity. Developments in analytical chemistry, much like the earlier revolution ushered in by next-generation DNA sequencing, are now transforming our understanding of these complex chemical communities. For example, a handful of lake water contains thousands of unique molecules of varying origin and composition, but we know little about how molecules change over time with their sources (e.g. terrestrial plants, aquatic microbes) and biogeochemical transformations. To compare ultra-high-resolution mass spectrometry (UHR-MS) data from different times and instruments, we need new statistical approaches that learn errors and correct the data, much like bioinformatics pipelines. Developing these tools will let us track DOM through time and space in ways that can inform ecosystem models.


Project Summary

In this project, you will develop new algorithms to process high throughput UHR-MS data and use these to analyse temporal dynamics of DOM at regional to global scales.  Once the algorithms are developed, you will choose two potential applications depending on your interests.  Applications include: i) predicting temporal dynamics of DOM at a global scale using monthly UHR-MS data from 70 lakes worldwide, some of which also have high-frequency sensor data; 
ii) determining the effects of forestry on the carbon balance of whole catchments using a repeat survey of DOM in 20 Swedish lake catchments where we have measured changes in surrounding land use and forest biomass; and iii) mapping DOM at high temporal resolution across landscapes by calibrating monthly airborne hyperspectral imagery with monthly UHR-MS data from 8 Canadian lake catchments.


What will the successful applicant do

You will join a large team working on the ERC-funded #sEEIngDOM project. In your first year, you will access hundreds of measurements of standards collected over a decade on different versions of a Fourier-transform ion cyclotron resonance mass spectrometer. You will also consider data on processing times, water chemistry of the parent matrix, and the efficiency with which DOM is separated from waters during analysis. You will develop a predictive machine learning algorithm to normalise standards run at a different time points to a known “gold” standard. You may work in the lab to develop an accompanying set of internal sample-specific standards, each with their own calibration curve, which could be used for normalising samples as is done with genetic sequencing data.  In later years, you will apply statistical and geospatial analyses to the existing datasets of interest with established pipelines in R, Python, GEE, working with external collaborators in Europe and North America. Samples for some application projects have also been collected but not yet measured with UHR-MS, so you will visit collaborators to collect these data and learn the latest techniques in chemoinformatics.


Training Provided

In addition to standard training in communication, project management, and experimental design, you will be trained in all lab work involving mass spectrometry (including data analysis) and statistical modelling (including computer programming). Further training in chemoinformatics, analytical chemistry, and machine learning will be provided through visits to European collaborators and formal courses, e.g. alignment with the AI4ER DTP.



  • Kellerman, A.M., Dittmar, T., Kothawala, D.N. & Tranvik, L.J. (2014) Chemodiversity of Dissolved Organic Matter in Lakes Driven by Climate and Hydrology. Nature Communications 5, 3804.
  • Hawkes, J. A., et al. (2020). An International Laboratory Comparison of Dissolved Organic Matter Composition by High Resolution Mass Spectrometry: Are We Getting the Same Answer? Limnology and Oceanography: Methods, 18 (6), 235–258.