patRoon: open source software platform for environmental mass spectrometry based non-target screening

Abstract

Mass spectrometry based non-target analysis is increasingly adopted in environmental sciences to screen and identify numerous chemicals simultaneously in highly complex samples. However, current data processing software either lack functionality for environmental sciences, solve only part of the workflow, are not openly available and/or are restricted in input data formats. In this paper we present patRoon, a new R based open-source software platform, which provides comprehensive, fully tailored and straightforward non-target analysis workflows. This platform makes the use, evaluation and mixing of well-tested algorithms seamless by harmonizing various common (primarily open) software tools under a consistent interface. In addition, patRoon offers various functionality and strategies to simplify and perform automated processing of complex (environmental) data effectively. patRoon implements several effective optimization strategies to significantly reduce computational times. The ability of patRoon to perform time-efficient and automated non-target data annotation of environmental samples is demonstrated with a simple and reproducible workflow using open-access data of spiked samples from a drinking water treatment plant study. In addition, the ability to easily use, combine and evaluate different algorithms was demonstrated for three commonly used feature finding algorithms. This article, combined with already published works, demonstrate that patRoon helps make comprehensive (environmental) non-target analysis readily accessible to a wider community of researchers.

Keywords: Compound identification; Computational workflows; High resolution mass spectrometry; Non-target analysis.

Fig. 1

Generic workflow for environmental non-target analysis

Fig. 2

Overview of the NTA patRoon workflow. All steps are optional. Steps that are connected by blue and straight arrows represent a one-way data dependency, whereas steps connected with red curved and dashed arrows represent steps with two-way data interaction

Fig. 3

Graphical user interface tools in patRoon. Tools are provided a to create a new patRoon data analysis project and b to inspect feature chromatography data

Fig. 4

Interface for the patRoon workflow. The workflow steps are performed by a set of functions that execute the selected algorithm and return the data in a harmonized format by utilizing the ‘S4’ object oriented programming approach of R. These objects all derive from a common base class and may be further sub-classed in algorithm specific classes (as is exemplified for features). Generic functions are defined for all workflow classes to implement further data processing functionality in a predictable and algorithm independent manner (see also Table 3). Further information is provided in the reference manual [85, 86]

Fig. 5

Parallelization benchmark results. a Benchmark results for commonly used CLI tools applied in patRoon workflows under varying parallelization conditions. The tested tools were msConvert, FeatureFinderMetabo (FFM), GenForm, SIRIUS and MetFrag. Tests were performed with “simple” (left) and “complex” (right) input conditions (Table 4) to simulate varying workflow complexity. Parallelization was performed with the multiprocessing functionality of patRoon (top) or by using native multithreading (bottom, for tools that supported this). Graphs represent number of processes or threads versus relative execution time (normalized to sequential results). The dotted grey lines represent the theoretical trend if maximum parallelization performance is achieved. The dashed blue line represents the number of physical cores that became the default selection in patRoon based on these results. b Comparison of execution times (normalized to the execution times of the unoptimized results) when tools are executed without optimizations (green), executed with native multithreading (FeatureFinderMetabo, SIRIUS and MetFrag) or batch mode (GenForm) (orange), executed with multiprocessing (purple) or a combination of the latter two (pink), using simple (left) and complex (right) input conditions. c Overview of execution times for a complete patRoon workflow executed under optimized versus unoptimized conditions. All results for msConvert and SIRIUS were obtained without enabling their native batch mode

Fig. 6

Common visualization functionality of patRoon applied to the demonstrated workflow. From left to right: an m/z vs retention time plot of all feature groups uniquely present in the samples, an EIC for the tramadol suspect, a compound annotated spectrum for the benzotriazole suspect and comparison of feature presence between sample groups using UpSet [77], Venn (influent/effluent A) and chord diagrams