Peak Lists Research Articles

Identifying Process Differences with ToF-SIMS: An MVA Decomposition Strategy.

In time-of-flight secondary ion mass spectrometry (ToF-SIMS), multivariate analysis (MVA) methods such as principal component analysis (PCA) are routinely employed to differentiate spectra. However, additional insights can often be gained by comparing processes, where each process is characterized by its own start and end spectra, such as when identical samples undergo slightly different treatments or when slightly different samples receive the same treatment. This study proposes a strategy to compare such processes by decomposing the loading vectors associated with them, which highlights differences in the relative behavior of the peaks. This strategy identifies key information beyond what is captured by the loading vectors or the end spectra alone. While PCA is widely used, partial least-squares discriminant analysis (PLS-DA) serves as a supervised alternative and is the preferred method for deriving process-related loading vectors when classes are narrowly separated. The effectiveness of the decomposition strategy is demonstrated using artificial spectra and applied to a ToF-SIMS materials science case study on the photodegradation of N719 dye, a common dye in photovoltaics, on a mesoporous TiO2 anode. The study revealed that the photodegradation process varies over time, and the resulting fragments have been identified accordingly. The proposed methodology, applicable to both labeled (supervised) and unlabeled (unsupervised) spectral data, can be seamlessly integrated into most modern mass spectrometry data analysis workflows to automatically generate a list of peaks whose relative behavior varies between two processes, and is particularly effective in identifying subtle differences between highly similar physicochemical processes.

A mass defect-based approach for the automatic construction of peak lists for databases of mass spectra with limited resolution: Application to time-of-flight secondary ion mass spectrometry data.

This study has developed a data processing protocol based on mass defect analysis for the automatic construction of unique peak lists addressing the need for the fast and efficient treatment of databases of mass spectra with limited mass resolution. The data processing protocol, implemented in MATLAB, is tested on a database of 126 mass spectra obtained from time-of-flight secondary ion mass spectrometry analysis of the exhaust of a laboratory diesel miniCAST burner deposited on Ti substrates. The data processing protocol converts the mass spectra into a data matrix suitable for chemometrics (peak list) by combining mass defect analysis and multivariate analysis. In particular, the role of the mass defect analysis is expanded to improve mass calibration and automate the construction of the peak list. In this context, mass defect analysis becomes an invaluable technique for the efficient processing of databases of mass spectra with limited mass resolution by allowing the fast and automated construction of a peak list common to all mass spectra, by improving the mass calibration, and finally by reducing the number of molecular formulae consistent with a given accurate mass, thus facilitating the identification of unknown ions.

Stereoselective Amine-omics Using Heavy Atom Isotope Labeled l-and d-Marfey's Reagents.

Biological amines and amino acids play essential roles in many biochemical processes. The chemical complexity of biological samples is challenging, and the selective identification and quantification of amines and amino acid stereoisomers would be very useful for amine-focused "amino-omics" studies. Many amines and amino acids are chiral, and their stereoisomers cannot be resolved on achiral media without chiral derivatization. In prior studies, we demonstrated the use of Marfey's reagent─a chiral derivatization reagent for amines and phenolic OH groups─for the LC-MS/MS resolution and quantification of amines and amino acid stereoisomers. In this study, a heavy atom isotope labeled Marfey's reagent approach for the stereoselective detection and quantification of amines and amino acids was developed. Heavy (13C2) l-Marfey's (Hl-Mar) and heavy (2H3) d-Marfey's (Hd-Mar) were synthesized from 13C2-l-Ala and 2H3-d-Ala, respectively. Both light and heavy Marfey's reagents were used to derivatize standard amine mixtures, which were analyzed by LC-QToF-HRMS. Aligned peak lists were comparatively analyzed by light vs heavy Mar mass differences to identify mono-, di-, and tri-Marfey's adducts and then by the retention time difference between l- and d-Mar derivatives to identify stereoisomers. This approach was then applied to identify achiral and chiral amine and amino acid components in a methicillin-resistant Staphylococcus aureus (MRSA) extract. This approach shows high analytical selectivity and reproducibility.

Lipid Wizard: Analysis Software for Comprehensive Two-Dimensional Liquid Chromatography-Mass Spectrometry-Based Lipid Profiling.

Lipids play a significant role in life activities and participate in the biological system through different pathways. Although comprehensive two-dimensional liquid chromatography-mass spectrometry (2DLC-MS) has been developed to profile lipid abundance changes, lipid identification and quantification from 2DLC-MS data remain a challenge. We created Lipid Wizard, open-source software for lipid assignment and isotopic peak stripping of the 2DLC-MS data. Lipid Wizard takes the peak list deconvoluted from the 2DLC-MS data as input and assigns each isotopic peak to the lipids recorded in the LIPID MAPS database by precursor ion m/z matching. The matched lipids are then filtered by the first-dimension retention time (1D RT), followed by the second-dimension retention time (2D RT), where the 2D RT of each lipid is predicted using an equivalent carbon number (ECN) model. The remaining assigned lipids are used for isotopic peak stripping via an iterative linear regression. The performance of Lipid Wizard was tested using a set of lipid standards and then applied to study the lipid changes in the livers of mice (fat-1) fed with alcohol.

ASAP: An automatic sequential assignment program for congested multidimensional solid state NMR spectra

Accurate signal assignments can be challenging for congested solid-state NMR (ssNMR) spectra. We describe an automatic sequential assignment program (ASAP) to partially overcome this challenge. ASAP takes three input files: the residue type assignments (RTAs) determined from the better-resolved NCACX spectrum, the full peak list of the NCOCX spectrum, and the protein sequence. It integrates our auto-residue type assignment strategy (ARTIST) with the Monte Carlo simulated annealing (MCSA) algorithm to overcome the hurdle for accurate signal assignments caused by incomplete side-chain resonances and spectral congestion. Combined, ASAP demonstrates robust performance and accelerates signal assignments of large proteins (>200 residues) that lack crystalline order.

Metabolite Profiling Foamy Substance of Cucumis sativus L. white local cultivar studied by HR-LCMS

Cucumis sativus L. belong to Cucurbitaceous. The Phytochemical profiling of white foamy substance was performed using HR-LCMS, details of projecting compounds with MS spectra, peak list, and compound structure were studied. The prominent constituent analysis (PCA) of the investigative data displayed the existence of Triterpenes: Gitoxigenin (7.76ppm), Strophanthidin(13.79ppm), Cucurbita in-C (-0.8ppm), Cucurbita in-E(5.53ppm), and Cardiac glycosides: and Digitoxigenin(-2.13ppm) in 10µl of the loaded sample.100 other composites were identified by accurate mass of Q-TOF/MS and confirmed by database compounds (IRM calibration) which comprises, abscisic acid amino acids, fatty acids pyridylactic acid etc. Strophanthidin in the premier attention in the white foamy extract is a noteworthy finding where it has interaction with various proteins in biochemical reactions. This study reports the customary implication of removing white foamy substances before slicing for ingestion. It is used as both vegetables as well as fruit in Indian traditional medicine. It is a source of bioactive compounds.

The 100-protein NMR spectra dataset: A resource for biomolecular NMR data analysis

Multidimensional NMR spectra are the basis for studying proteins by NMR spectroscopy and crucial for the development and evaluation of methods for biomolecular NMR data analysis. Nevertheless, in contrast to derived data such as chemical shift assignments in the BMRB and protein structures in the PDB databases, this primary data is in general not publicly archived. To change this unsatisfactory situation, we present a standardized set of solution NMR data comprising 1329 2–4-dimensional NMR spectra and associated reference (chemical shift assignments, structures) and derived (peak lists, restraints for structure calculation, etc.) annotations. With the 100-protein NMR spectra dataset that was originally compiled for the development of the ARTINA deep learning-based spectra analysis method, 100 protein structures can be reproduced from their original experimental data. The 100-protein NMR spectra dataset is expected to help the development of computational methods for NMR spectroscopy, in particular machine learning approaches, and enable consistent and objective comparisons of these methods.

Magnetstein: An Open-Source Tool for Quantitative NMR Mixture Analysis Robust to Low Resolution, Distorted Lineshapes, and Peak Shifts.

1H NMR spectroscopy is a powerful tool for analyzing mixtures including determining the concentrations of individual components. When signals from multiple compounds overlap, this task requires computational solutions. They are typically based on peak-picking and the comparison of obtained peak lists with libraries of individual components. This can fail if peaks are not sufficiently resolved or when peak positions differ between the library and the mixture. In this paper, we present Magnetstein, a quantification algorithm rooted in the optimal transport theory that makes it robust to unexpected frequency shifts and overlapping signals. Thanks to this, Magnetstein can quantitatively analyze difficult spectra with the estimation trueness an order of magnitude higher than that of commercial tools. Furthermore, the method is easier to use than other approaches, having only two parameters with default values applicable to a broad range of experiments and requiring little to no preprocessing of the spectra.

NECTAR: A New Algorithm for Characterizing and Correcting Noise in QToF-Mass Spectrometry Imaging Data.

A typical mass spectrometry imaging experiment yields a very high number of detected peaks, many of which are noise and thus unwanted. To select only peaks of interest, data preprocessing tasks are applied to raw data. A statistical study to characterize three types of noise in MSI QToF data (random, chemical, and background noise) is presented through NECTAR, a new NoisE CorrecTion AlgoRithm. Random noise is confirmed to be dominant at lower m/z values (∼50-400 Da) while systematic chemical noise dominates at higher m/z values (>400 Da). A statistical approach is presented to demonstrate that chemical noise can be corrected to reduce its presence by a factor of ∼3. Reducing this effect helps to determine a more reliable baseline in the spectrum and therefore a more reliable noise level. Peaks are classified according to their spatial S/N on the single ion images, and background noise is thus removed from the list of peaks of interest. This new algorithm was applied to MALDI and DESI QToF data generated from the analysis of a mouse pancreatic tissue section to demonstrate its applicability and ability to filter out these types of noise in a relevant data set. PCA and t-SNE multivariate analysis reviews of the top 4000 peaks and the final 744 and 299 denoised peak list for MALDI and DESI, respectively, suggests an effective removal of uninformative peaks and proper selection of relevant peaks.

XRD analysis of nanoparticles synthesized using aqueous and alcoholic extracts of Cuscuta reflexa

Plants are one of the most abundant sources of biomolecules on the planet. Researchers and academics have become more interested in the antibacterial and therapeutic characteristics of plants over the last decade. The introduction and combination of nanotechnology and phytochemistry brought up new possibilities. The synthesis of metallic nanoparticles mediated by plant extracts developed quickly and has been extensively researched. In this paper, we describe the manufacture of silver nanoparticles from Cuscuta reflexa aqueous and alcoholic extracts. For characterisation, the nanoparticles were subjected to X-ray diffraction (XRD) examination. XRD analysis offers concrete information about the structure and crystalline size of nanoparticles, and it can play an important role in nanoparticle characterisation. The observation includes findings for matched phases, search-match, selection criteria, peak list, crystallite size estimation, crystallinity analysis, diffraction pattern graphics, and so on. The nano size of the crystals was disclosed by XRD examination, which proved to be 38.55 nm and 66.27 nm in the case of nanoparticles synthesised using aqueous and alcoholic extracts of Cuscuta reflexa, respectively. The crystal system was reported to be cubic with side length of 0.4 nm, the calculated density of silver in nanoparticles was calculated to be 10.506 g/cm3 which corresponds with the density of silver element, i.e., 10.49 g/cm3

MultiplexMS: A Mass Spectrometry-Based Multiplexing Strategy for Ultra-High-Throughput Analysis of Complex Mixtures.

High-throughput chemical analysis of natural product mixtures lags behind developments in genome sequencing technologies and laboratory automation, leading to a disconnect between library-scale chemical and biological profiling that limits new molecule discovery. Here, we report a new orthogonal sample multiplexing strategy that can increase mass spectrometry-based profiling up to 30-fold over traditional methods. Profiled pooled samples undergo subsequent computational deconvolution to reconstruct peak lists for each sample in the set. We validated this approach using in silico experiments and demonstrated a high assignment precision (>97%) for large, pooled samples (r = 30), particularly for infrequently occurring metabolites of relevance in drug discovery applications. Requiring only 5% of the previously required MS acquisition time, this approach was repeated in a recent biological activity profiling study on 925 natural product extracts, leading to the rediscovery of all previously reported bioactive metabolites. This new method is compatible with MS data from any instrument vendor and is supported by an open-source software package: https://github.com/liningtonlab/MultiplexMS.

Blind assessment of monomeric AlphaFold2 protein structure models with experimental NMR data

Recent advances in molecular modeling of protein structures are changing the field of structural biology. AlphaFold-2 (AF2), an AI system developed by DeepMind, Inc., utilizes attention-based deep learning to predict models of protein structures with high accuracy relative to structures determined by X-ray crystallography and cryo-electron microscopy (cryoEM). Comparing AF2 models to structures determined using solution NMR data, both high similarities and distinct differences have been observed. Since AF2 was trained on X-ray crystal and cryoEM structures, we assessed how accurately AF2 can model small, monomeric, solution protein NMR structures which (i) were not used in the AF2 training data set, and (ii) did not have homologous structures in the Protein Data Bank at the time of AF2 training. We identified nine open-source protein NMR data sets for such “blind” targets, including chemical shift, raw NMR FID data, NOESY peak lists, and (for 1 case) 15N-1H residual dipolar coupling data. For these nine small (70–108 residues) monomeric proteins, we generated AF2 prediction models and assessed how well these models fit to these experimental NMR data, using several well-established NMR structure validation tools. In most of these cases, the AF2 models fit the NMR data nearly as well, or sometimes better than, the corresponding NMR structure models previously deposited in the Protein Data Bank. These results provide benchmark NMR data for assessing new NMR data analysis and protein structure prediction methods. They also document the potential for using AF2 as a guiding tool in protein NMR data analysis, and more generally for hypothesis generation in structural biology research.

Abstract 3504: Evaluating the molecular determinants of PDAC racial health disparities

Abstract Pancreatic cancer is the 3rd leading cause of cancer-related deaths with a 5-year survival rate of just 11.5%. Pancreatic ductal adenocarcinoma (PDAC), accounting for over 90% of pancreatic malignancies, has the highest incidence rate, mortality rate, and shortest survival times in non-Hispanic black (defined here as African and/or African American ancestry) patients compared to other races. Due to the racial health disparities of PDAC, this study seeks to quantify the proteomic signatures associated with tumor racial origin to identify the underlying molecular pathways that could contribute to these disparities. Using mass spectrometry on primary PDAC tumor samples collected from 30 Caucasian and 12 African American patients, 183 proteins were differentially expressed between these groups. The most over-expressed protein in African American tumors was Ring Finger Protein 2 (RNF2) with a log2 fold-change greater than six. RNF2 had previously been found to regulate GATA Binding Protein 6 (GATA6) in embryonic stem cells. Additionally, GATA6 is also an important determinant in PDAC subtyping with higher GATA6 expression leading to greater chemosensitivity. To expand upon these findings, chromatin immunoprecipitation and sequencing (ChIP-seq) was performed on MIA-PaCa2 to identify binding sites for RNF2 in PDAC cells. Peaks were called with MACS2 (q <0.01) and a final peak list with an irreproducible discovery rate (IDR) of less than 0.05 across two replicates was generated. From the 1,142 replicated peaks identified, RNF2 peaks were localized over the transcription start site of 644 unique transcripts, including GATA6. Modulating RNF2 levels by exogenous expression or short hairpin RNA (shRNA) knockdown confirms that GATA6 is a target of RNF2 repression in PDAC. To further identify its downstream targets in PDAC, RNF2 knock down in MIA-PaCa2 was analyzed by RNA-seq. Differential gene expression analysis revealed RNF2 knockdown decreased expression of genes related to innate immunity, which we have previously identified as a marker of the aggressive Inflammatory PDAC subtype using proteomics. Of the 4,763 differentially expressed genes (P-adj < 0.05), 226 had RNF2 binding sites, demonstrating the scope of RNF2 regulation in MIA-PaCa2. Our work identifies differences in the underlying proteome of PDAC tumors as a function of racial origin, with overexpression of RNF2 in African American tumors potentially contributing to inflammation and subtype delineation. Because PDAC subtypes are predictive of therapeutic response, additional research is necessary to determine whether targeting RNF2 can mitigate oncogenic phenotypes that contribute to racial health disparities. Citation Format: Anthony E. Johansen-Sallee, Alexa M. Barber, Henry Law, Jose Trevino, Nicholas T. Woods. Evaluating the molecular determinants of PDAC racial health disparities [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3504.

MALDI Imaging, a Powerful Multiplex Approach to Decipher Intratumoral Heterogeneity: Combined Hepato-Cholangiocarcinomas as Proof of Concept

Combined hepato-cholangiocarcinomas (cHCC-CCA) belong to the spectrum of primary liver carcinomas, which include hepatocellular carcinomas (HCC) and intrahepatic cholangiocarcinomas (iCCA) at both ends of the spectrum. Mainly due to the high intratumor heterogeneity of cHCC-CCA, its diagnosis and pathological description remain challenging. Taking advantage of in situ non-targeted molecular mapping provided by MALDI (Matrix Assisted Laser Desorption Ionization) imaging, we sought to develop a multiscale and multiparametric morphological approach, integrating molecular and conventional pathological analysis. MALDI imaging was applied to five representative cases of resected cHCC-CCA. Principal component analysis and segmentations with MALDI imaging techniques identified areas related to either iCCA or HCC and also hidden tumor areas not visible microscopically. In addition, the overlap between MALDI segmentation and immunostaining provided a comprehensive description of cHCC-CCA tumor heterogeneity by identifying transitional and micro-metastatic areas. Moreover, a list of peptides derived from in silico digestion was obtained for each immunohistochemical marker and was matched within the peptide peak list acquired by MALDI. Comparison of immunostaining images with ions from in silico digestion revealed an accurate identification of iCCA and HCC areas. Our study provides further evidence on the performance of MALDI imaging in exploring intratumor heterogeneity and offering virtual multiplex immunostaining through a single acquisition.

A Robust Purity Method for Biotherapeutics Using New Peak Detection in an LC–MS-Based Multi-Attribute Method

New peak detection (NPD), as part of the LC-MS-based multi-attribute method (MAM), allows for sensitive and unbiased detection of new or changing site-specific attributes between a sample and reference that is not possible with conventional UV or fluorescence detection-based methods. MAM with NPD can serve as a purity test that can establish whether a sample and the reference are similar. The broad implementation of NPD in the biopharmaceutical industry has been limited by the potential presence of false positives or artifacts, which increase the analysis time and can trigger unnecessary investigations of product quality. Our novel contributions to the success of NPD are the curation of false positives, use of the known peak list concept, pairwise analysis approach, and the development of a NPD system suitability control strategy. In this report, we also introduce a unique experimental design utilizing sequence variant co-mixes to measure NPD performance. We show that NPD has superior performance relative to conventional control system methods in the detection of an unexpected change as compared with the reference. NPD is a new frontier in purity testing that reduces subjectivity, need for analyst intervention, and potential for missing unexpected product quality changes.

SsPINE/ssPINE-POKY: Automated chemical shift assignment with an intuitive graphical user interface for solid-state NMR data from complex protein systems.

In structural biology, nuclear magnetic resonance (NMR) is an important method when studying biological complexes with high resolution. Compared with solution NMR, solid-state NMR (ssNMR) has increased the possibility of studying large macromolecular assemblies with high structural complexity mainly related to low solubility. However, computational tools for the analysis of complex multidimensional ssNMR data have lagged behind those for solution NMR. Before a structure can be determined, thousands of signals from individual types of multidimensional ssNMR spectra of samples must be recognized, correlated, categorized, and eventually assigned to atoms in the chemical structure. To address these tedious steps, we have developed an automated algorithm for ssNMR spectra called “ssPINE”. The ssPINE software accepts the sequence of the protein plus peak lists from a variety of ssNMR experiments as inputs and offers automated backbone and side-chain assignments. To lower the bar of using ssPINE, we have developed a graphical user interface called ssPINE-POKY as a plugin to POKY that provides seamless communication between POKY and ssPINE webserver. The user only needs to select experiments in POKY to run a job and import results with just a few mouse clicks. Supported by NSF DBI-2051595, DBI-1902076 and University of Colorado Denver.

Phytochemical analysis in foamy extract of cucumber (Cucumis sativus L.)

The metabolite profiling of white foamy substance local white cultivar of Cucumber (Cucumis sativus L.) was performed using HR-LCMS. Prominent compounds with MS spectra, peak list, and compound structure were studied. Which pertains to Cucurbitaceae. The cytotoxic activity was studied on onion root tip squash and microscopic evaluation was done. The principal component analysis (PCA) of the analytical data showed the presence of Triterpenes: Cucurbitacin-C (0.8ppm), Cucurbitacin-E (5.53ppm) and Cardiac glycosides: Gitoxigenin (7.76ppm), Strophanthidin (13.79ppm), and Digitoxigenin (2.13ppm) in 10μl of the loaded sample.100 other compounds were identified by accurate mass of Q-TOF/MS and verified by database compounds (IRM calibration). Strophanthidin in the highest concentration in the foamy extract is a significant finding. This study reports the traditional significance of removing white foamy substances for bitterness in cucumber before slicing. The cytotoxic study on onion root tips by photomicrograph study indicates reduction in root growth and chromosomal aberrations, disturbed telophase, anaphase, abnormal cell shape, dissolved chromosomes, membrane damage and nuclear lesions. SAARC J. Agric., 20(2): 145-155 (2022)

Deconvolution of 1D NMR spectra: A deep learning-based approach

The analysis of nuclear magnetic resonance (NMR) spectra to detect peaks and characterize their parameters, often referred to as deconvolution, is a crucial step in the quantification, elucidation, and verification of the structure of molecular systems. However, deconvolution of 1D NMR spectra is a challenge for both experts and machines. We propose a robust, expert-level quality deep learning-based deconvolution algorithm for 1D experimental NMR spectra. The algorithm is based on a neural network trained on synthetic spectra. Our customized pre-processing and labeling of the synthetic spectra enable the estimation of critical peak parameters. Furthermore, the neural network model transfers well to the experimental spectra and demonstrates low fitting errors and sparse peak lists in challenging scenarios such as crowded, high dynamic range, shoulder peak regions as well as broad peaks. We demonstrate in challenging spectra that the proposed algorithm is superior to expert results.

Abstract PR010: Characterizing N-glycan profiles of DCIS progression using tissue imaging MALDI mass spectrometry

Abstract Identifying distinct biomarkers of DCIS and determining the invasive potential of DCIS are of great clinical importance. N-linked glycans present on the cell surface of DCIS lesions and surrounding stroma are potential biomarkers of disease and tissue, but remain largely unexplored. An N-glycan targeted imaging mass spectrometry (IMS) approach was initiated to identify N-glycans associated with DCIS and progression in clinical FFPE tissues. A cohort of pathologist annotated DCIS and IDC tissues with a range of disease severity as well as the RAHBT TMA DCIS cohort were evaluated. The RAHBT cohort contains primary DCIS lesions with known long-term outcomes and serves as a good sample set to identify early markers indicating progressive potential, while the DCIS and IDC large tissue biopsies provide information on distinct glycan profiles when DCIS has progressed to IDC. Initially, tissue samples containing DCIS alone or both DCIS and IDC were processed for N-glycan MALDI-IMS analysis, the goal being to establish a consensus panel of each analyte and determine distinct profiles of DCIS when alone vs. when IDC was present. For N-glycans, a peak list of 54 N-glycans was selected for evaluation of each DCIS only tissue, as well as the mixed DCIS/IDC tissues. By highest intensity, the major DCIS-associated glycans were in the high-mannose and paucimannose structure categories. Additionally, a series of tri- and tetra-antennary multi-fucosylated glycans were also detected specifically in the DCIS lesions. In tissues containing both IDC and DCIS lesions, the high mannose glycans were also detected in the IDC regions, and GlcNAc-bisected glycans were seen elevated in these as compared to DCIS-only tissues. For samples in the RAHBT TMA cohort (progressors N=43; non-progressors N=78), the bisected N-glycans seen elevated with IDC were also seen to be significantly increased in the RAHBT samples that would eventually progress to IDC. Additional analyses to evaluate N-glycan isomer distributions for fucosylated and sialylated glycan species are also ongoing. The cumulative N-glycan data will be assessed with the extensive spatial genomic, transcriptomic and immunohistological data already generated for the same samples in the RAHBT cohort. This novel combination of multi-enzymatic digests with histopathology annotations represents an extensive multi-dimensional profile of DCIS and a novel tissue biomarker discovery approach. Citation Format: Elizabeth N. Wallace, Grace Grimsley, Siri H. Strand, Robert Michael Angelo, Graham Colditz, E. Shelley Hwang, Robert West, Jeffrey R. Marks, Peggi M. Angel, Richard R. Drake. Characterizing N-glycan profiles of DCIS progression using tissue imaging MALDI mass spectrometry [abstract]. In: Proceedings of the AACR Special Conference on Rethinking DCIS: An Opportunity for Prevention?; 2022 Sep 8-11; Philadelphia, PA. Philadelphia (PA): AACR; Can Prev Res 2022;15(12 Suppl_1): Abstract nr PR010.

Deep denoising autoencoder-assisted continuous scoring of peak quality in high-resolution LC−MS data

Accurate peak picking is a challenging but fundamental problem in LC-MS-based omics analysis. Previous efforts mainly focused on continuous wavelet transform (CWT) based peak detection, which is highly sensitive while suffers from unsatisfactory precision. Recently proposed deep learning (DL) based peak classifiers improve the performance significantly. However, their classification strategy loses the continuous criterion for controlling the false positive rate flexibly. Here we put forward AutoMS, which employs a deep learning-based denoising autoencoder to grasp the common characteristics of chromatographic peaks, and predict noise-deducted peaks from the original peak profiles. By comparing the difference before and after processed, it scores the peak quality continuously and precisely. From the evaluating result, AutoMS improved the accuracy for peak picking. AutoMS integrates HPIC for ROI extraction in order to accept raw data directly and output quantitative results. It also supports peak lists obtained from other tools with little adjustment. AutoMS is open source and available at https://github.com/hcji/AutoMS.