Biology subjects

Seidl, B.

Publications and source records attributed to Seidl, B..

3 recordsLinked to original sources

bioRxiv2022

Heterogeneous multimeric metabolite ion species observed in LC-MS based metabolomics data sets

Covalent or non-covalent heterogeneous multimerization of molecules associated with extracts from biological samples analyzed via LC-MS is quite difficult to recognize/annotate and therefore the prevalence of multimerization remains largely unknown. In this study, we utilized 13C labeled and unlabeled Pichia pastoris extracts to recognize heterogeneous multimers. More specifically, between 0.8% and 1.5% of the biologically-derived features detected in our experiments were confirmed to be heteromers, about half of which we could successfully annotate with monomeric partners. Interestingly, we found specific chemical classes such as nucleotides to disproportionately contribute to heteroadducts. Furthermore, we compiled these compounds into the first MS/MS library that included data from heteromultimers to provide a starting point for other labs to improve the annotation of such ions in other metabolomics data sets. Then, the detected heteromers were also searched in publicly accessible LC-MS datasets available in Metabolights, Metabolomics WB, and GNPS/MassIVE to demonstrate that these newly annotated ions are also relevant to other public datasets. Furthermore, in additional datasets (Triticum aestivum, Fusarium graminearum, and Trichoderma reesei) our developed workflow also detected 0.5% to 4.9% of metabolite features to originate from heterodimers, demonstrating heteroadducts to be present in metabolomics studies at a low percentage.

systems biology↗

bioRxiv2021

CPExtract, a Software for the Automated Tracer-Based Pathway Specific Screening of Secondary Metabolites in LC-HRMS Data

The use of stable isotopically labeled tracers is a long-proven way of specifically detecting and tracking derived metabolites through a metabolic network of interest. While recently developed stable isotope assisted methods and associated, supporting data analysis tools have greatly improved untargeted metabolomics approaches, no software tool is currently available that allows to automatically search LC-HRMS chromatograms for completely free user-definable isotopolog patterns expected for the metabolism of labeled tracer substances. Here we present Custom Pattern Extract (CPExtract), a versatile software tool that allows for the first time the high-through-put search for user-defined isotopolog patterns in LC-HRMS data. The patterns can be specified via a set of rules including the presence or absence of certain isotopologs, their relative intensity ratios as well as chromatographic co-elution. Each isotopolog pattern satisfying the respective rules is verified on a MS-scan level and also in the chromatographic domain. The CPExtract algorithm allows the use of both labeled tracer compounds in non-labeled biological samples as well as a reversed tracer approach, employing non-labeled tracer compounds along with globally labeled biological samples. In a proof of concept study we searched for metabolites specifically arising from the malonate pathway of the filamentous fungi Fusarium graminearum and Trichoderma reesei. 1,2,3-13C3-malonic acid diethyl ester and native malonic acid monomethyl ester were used as tracers. We were able to reliably detect expected fatty acids and known polyketides. In addition, up to 189 and 270 further, unknown metabolites presumably including novel polyketides were detected in the F. graminearum and T. reesei culture samples respectively, all of which exhibited the user-predicted isotopolog patterns originating from the malonate tracer incorporation. The software can be used for every conceivable tracer approach. Furthermore, the rule sets can be easily adapted or extended if necessary. CPExtract is available free of charge for non-commercial use at https://metabolomics-ifa.boku.ac.at/CPExtract.

bioinformatics↗

bioRxiv2021

PeakBot: Machine learning based chromatographic peak picking

MotivationChromatographic peak picking is among the first steps in data processing workflows of raw LC-HRMS datasets in untargeted metabolomics applications. Its performance is crucial for the holistic detection of all metabolic features as well as their relative quantification for statistical analysis and metabolite identification. Random noise, non-baseline separated compounds and unspecific background signals complicate this task. ResultsA machine-learning framework entitled PeakBot was developed for detecting chromatographic peaks in LC-HRMS profile-mode data. It first detects all local signal maxima in a chromatogram, which are then extracted as super-sampled standardized areas (retention-time vs. m/z). These are subsequently inspected by a custom-trained convolutional neural network that forms the basis of PeakBots architecture. The model reports if the respective local maximum is the apex of a chromatographic peak or not as well as its peak center and bounding box. In training and independent validation datasets used for development, PeakBot achieved a high performance with respect to discriminating between chromatographic peaks and background signals (accuracy of 0.99). For training the machine-learning model a minimum of 100 reference features are needed to learn their characteristics to achieve high-quality peak-picking results for detecting such chromatographic peaks in an untargeted fashion. PeakBot is implemented in python (3.8) and uses the TensorFlow (2.5.0) package for machine-learning related tasks. It has been tested on Linux and Windows OSs. AvailabilityThe package is available free of charge for non-commercial use (CC BY-NC-SA). It is available at https://github.com/christophuv/PeakBot. Contactchristoph.bueschl@univie.ac.at Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗