Graph Theoretic Approach for the Analysis of Comprehensive Mass-Spectrometry (MS/MS) Data of Dissolved Organic Matter

Abstract

Dissolved organic matter (DOM) is a highly complex mixture of organic substances found in aquatic ecosystems. This mixture results from the degradation of primary producers within the ecosystem, groundwater, and the surrounding terrestrial sources. Understanding the chemical structure of DOM is crucial to assessing its impact on aquatic ecosystems. Although multiple studies have addressed the complexity of DOM, the molecular structure of this set of compounds remains unclear. In this work, we present a novel computational framework "Graph-DOM," to assess the comprehensive fragmentation data obtained from the analysis of DOM using the Data Independent Fragmentation strategy with ESI-FT-ICR MS/MS enabling better understanding of the structural complexity of DOM. Graph-DOM uses graph algorithms to dissect a compiled output file obtained from processing hundreds of ultra-high-resolution fragment spectra. Over half a million ordered fragmentation pathways were computed for 764 isolated precursor ions assuming up to seven vector segments categorized as neutral losses (CH2, CH3, O, CH4, H2O, CO, and CO2). Families of structurally related molecules were identified using pathway overlaps, and output files compatible with network visualization software (e.g., Cytoscape) were also generated. Graph-DOM is able to efficiently process all the pathways to discover families within only a few minutes with adjustable parameters for overlap length of fragmentation pathways as well as configuring low abundance CHOS, CHON, and CHONS compounds. Graph-DOM is available at https://github.com/Usman095/Graph-DOM.

Keywords: HPC; dissolved organic matter; dynamic programming; regularity chains.

Fig. 1.

Sample spectrum of nominal mass 391 Da. The spectrum is analyzed by selecting a precursor ion at the nominal mass (±1 Da) here [C₁₉H₂₀O₉]⁻ and then constructing a directed acyclic graph (DAG) using all the neutral losses. Above, three possible pathways are shown with two (red) leading to core-fragment [C₁₀H₉O₂]⁻ and one (blue) leading to core-fragment [C₁₂H₁₁O]⁻. As can be seen, the paths can overlap, and multiple paths can lead to a single core-fragments generating a DAG as shown in figure 2.

Fig. 2.

DAG corresponding to the spectrum in figure 1. The branches are constructed by adding neutral losses such that a fragment of m/z of precursor minus the neutral loss is found in the spectrum. The DAG is grown recursively and splitting/joining branches wherever the paths overlap or split. As can be seen, two paths (red) start out separate but then join at the last neutral loss resulting in a common core fragment. Multiple fragmentation pathways leading to the same core fragment provide evidence for the presence of structural isomers. On the other hand (blue) might only have one branch leading to a unique core fragment.

Fig. 3.

Pathways constructed by traversing the DAG from the root (precursor) node to all the leaf (core-fragment) nodes. Note that each unique pathway is listed separately as the intermediate fragment sequence is important for identifying related families using fragmentation pathway overlap.

Fig. 4.

Pathways that will be connected by edges in the Family DAG. As can be seen, for every larger precursor, the next smaller precursor is the same as its first intermediate fragment. Matching precursors and intermediate fragments are highlighted with the same color.

Fig. 5.

Family DAG constructed from identified pathways. Each node represents a pathway while a path from the root node to a leaf node comprises a family.

Fig. 6.

Distribution of number of pathway w.r.t. to the precursor m/z. As shown in the plot, depending on the mass-spectrometer settings, and the chemical properties of the molecules, number of fragmentation pathways can vary vastly up to more than 40k per sample.

Fig. 7.

Distribution of number of core-fragments w.r.t. to the precursor m/z.

Fig. 8.

Distribution of number of families w.r.t the family size and overlap length used to identify families.

Fig. 9.

Ordered distribution of confidence value for each unique family.