Development of Bioinformatic Methods to Determine Chemical Differences Between Nematode Species

The metabolome is the complete set of small-molecule chemicals that are produced during metabolism and is an indicator of cellular activity and the physiological state of an organism. Understanding the complexities and impact of an organism’s complete metabolome, as well as variation between organisms has proven challenging. However, by studying the complete anatomy of an organism, whilst preserving biological complexity, can dramatically improve the amenability to analysis by studying smaller and simpler model organisms such as nematodes, e.g., Caenorhabditis elegans and Pristionchus pacificus. Furthermore, recent technological advances in surface sensitive mass spectrometry, such as the 3D-OrbiSIMS, have made it possible to not only characterise the entire metabolome of an organism, but to also spatially locate where metabolites anatomically originate, to a high spatial and mass accuracy. Therefore, the overarching aim of this study is to develop apply and evaluate bioinformatic methods for 3D-OrbiSIMS metabolome datasets for nematodes C. elegans and P. pacificus to separate species based on their mass spectra. Nematode samples were synchronised to larval stages and were deposited on indium tin oxide (ITO) substrates, to optimise charge compensation and were imaged in negative and positive polarities, as well as in XY axis & Z axis to obtain surface and depth chemical maps. Open source pySPM written in python was evaluated to determine its utility to analyse 3D-OrbiSIMS data. However, it was soon established that file formats produced by 3D-OrbiSIMS were incompatible. Therefore, 3D-OribSIMS native SurfaceLab v7 was used to interpret and extract data. SurfaceLab indicated the nematode mass spectra were extremely rich in chemical information with >250,000 mass ions for a number of data sets. Compromises to the large data sets, whilst preserving data complexity, were established by capturing the 5000 highest intensity mass ions for each sample and merging into master mass interval list which consisted of 24,667 mass ions to overcome processing limitations. The master mass interval list permitted multivariate principal component and principal coordinate analysis (PCA & PCoA) and unsupervised analysis (K-means and hierarchical clustering) of data to identify global similarities and differences between species which was conducted using R. PCA and PcoA were unable to separate C. elegans and P. pacificus data. PC1, PC2 and PC3, consisted of the greatest variance in the data. However, no 2D principal component plot (PC1 vs PC2, PC1 vs PC3 and PC2 vs PC3) were able to separate species mass spectra, when visualised using 95% confidence intervals. This was attributed to the large variance in the number of nematodes per sample (n = 1-59 nematodes) and number of data sets for between XY axis data (n=1-3) and Z axis data (n = 4-6). 3D variance plots for PC1, PC2 and PC3 appeared to separate the species. PcoA, like PCA, was also unable to separate the data. However, exploration of the array of distance matrix calculations (Euclidian, Manhattan Maximum and Canberra) identified that Canberra distances, due to its sensitivity to values close to zero, demonstrated potential to separate species, although it did not achieve this with 95% confidence. Chemistries were grouped using secondary ion mass spectrometry molecular formula prediction (SIMSMFP) to determine if biological classes were able to differentiate between species based on mass spectrometry data. Chemistries were grouped, for substrate (silica glass and ITO), fatty acids, triglycerides sulphates, ceramides, phospholipids, amino acids and nucleic bases. Fatty acids (1.54% of data) and triglycerides (3.54% of data) were the only chemical subgroups that demonstrated 90% success in chemically separating C. elegans and P. pacificus based on predicted values using K-means analysis, dendrograms, cluster plots and heatmaps. Dendrogram’s correctly separated all but one sample The data anomaly reasons are to be determined, however, could be attributed to sample contamination. In summary, the observations provide evidence to suggest holistic analyses of nematode chemistry is challenging as the C. elegans and P. pacificus are potentially more similar to each other than they are different. The research conducted as part of this study will pave the way towards building a WOrMs platform, which could be used to determine the spatial origin of ageing for free living and parasitic nematodes, that could improve the understanding of all forms of life